↑ Collapse diff ↑
Ignore white space 6 line context
1
/* -*- C++ -*-
2
 *
3
 * This file is a part of LEMON, a generic C++ optimization library
4
 *
5
 * Copyright (C) 2003-2008
6
 * Egervary Jeno Kombinatorikus Optimalizalasi Kutatocsoport
7
 * (Egervary Research Group on Combinatorial Optimization, EGRES).
8
 *
9
 * Permission to use, modify and distribute this software is granted
10
 * provided that this copyright notice appears in all copies. For
11
 * precise terms see the accompanying LICENSE file.
12
 *
13
 * This software is provided "AS IS" with no warranty of any kind,
14
 * express or implied, and with no claim as to its suitability for any
15
 * purpose.
16
 *
17
 */
18

	
19
#ifndef LEMON_BELLMAN_FORD_H
20
#define LEMON_BELLMAN_FORD_H
21

	
22
/// \ingroup shortest_path
23
/// \file
24
/// \brief Bellman-Ford algorithm.
25

	
26
#include <lemon/bits/path_dump.h>
27
#include <lemon/core.h>
28
#include <lemon/error.h>
29
#include <lemon/maps.h>
30
#include <lemon/path.h>
31

	
32
#include <limits>
33

	
34
namespace lemon {
35

	
36
  /// \brief Default OperationTraits for the BellmanFord algorithm class.
37
  ///  
38
  /// This operation traits class defines all computational operations
39
  /// and constants that are used in the Bellman-Ford algorithm.
40
  /// The default implementation is based on the \c numeric_limits class.
41
  /// If the numeric type does not have infinity value, then the maximum
42
  /// value is used as extremal infinity value.
43
  template <
44
    typename V, 
45
    bool has_inf = std::numeric_limits<V>::has_infinity>
46
  struct BellmanFordDefaultOperationTraits {
47
    /// \e
48
    typedef V Value;
49
    /// \brief Gives back the zero value of the type.
50
    static Value zero() {
51
      return static_cast<Value>(0);
52
    }
53
    /// \brief Gives back the positive infinity value of the type.
54
    static Value infinity() {
55
      return std::numeric_limits<Value>::infinity();
56
    }
57
    /// \brief Gives back the sum of the given two elements.
58
    static Value plus(const Value& left, const Value& right) {
59
      return left + right;
60
    }
61
    /// \brief Gives back \c true only if the first value is less than
62
    /// the second.
63
    static bool less(const Value& left, const Value& right) {
64
      return left < right;
65
    }
66
  };
67

	
68
  template <typename V>
69
  struct BellmanFordDefaultOperationTraits<V, false> {
70
    typedef V Value;
71
    static Value zero() {
72
      return static_cast<Value>(0);
73
    }
74
    static Value infinity() {
75
      return std::numeric_limits<Value>::max();
76
    }
77
    static Value plus(const Value& left, const Value& right) {
78
      if (left == infinity() || right == infinity()) return infinity();
79
      return left + right;
80
    }
81
    static bool less(const Value& left, const Value& right) {
82
      return left < right;
83
    }
84
  };
85
  
86
  /// \brief Default traits class of BellmanFord class.
87
  ///
88
  /// Default traits class of BellmanFord class.
89
  /// \param GR The type of the digraph.
90
  /// \param LEN The type of the length map.
91
  template<typename GR, typename LEN>
92
  struct BellmanFordDefaultTraits {
93
    /// The type of the digraph the algorithm runs on. 
94
    typedef GR Digraph;
95

	
96
    /// \brief The type of the map that stores the arc lengths.
97
    ///
98
    /// The type of the map that stores the arc lengths.
99
    /// It must conform to the \ref concepts::ReadMap "ReadMap" concept.
100
    typedef LEN LengthMap;
101

	
102
    /// The type of the arc lengths.
103
    typedef typename LEN::Value Value;
104

	
105
    /// \brief Operation traits for Bellman-Ford algorithm.
106
    ///
107
    /// It defines the used operations and the infinity value for the
108
    /// given \c Value type.
109
    /// \see BellmanFordDefaultOperationTraits
110
    typedef BellmanFordDefaultOperationTraits<Value> OperationTraits;
111
 
112
    /// \brief The type of the map that stores the last arcs of the 
113
    /// shortest paths.
114
    /// 
115
    /// The type of the map that stores the last
116
    /// arcs of the shortest paths.
117
    /// It must conform to the \ref concepts::WriteMap "WriteMap" concept.
118
    typedef typename GR::template NodeMap<typename GR::Arc> PredMap;
119

	
120
    /// \brief Instantiates a \c PredMap.
121
    /// 
122
    /// This function instantiates a \ref PredMap. 
123
    /// \param g is the digraph to which we would like to define the
124
    /// \ref PredMap.
125
    static PredMap *createPredMap(const GR& g) {
126
      return new PredMap(g);
127
    }
128

	
129
    /// \brief The type of the map that stores the distances of the nodes.
130
    ///
131
    /// The type of the map that stores the distances of the nodes.
132
    /// It must conform to the \ref concepts::WriteMap "WriteMap" concept.
133
    typedef typename GR::template NodeMap<typename LEN::Value> DistMap;
134

	
135
    /// \brief Instantiates a \c DistMap.
136
    ///
137
    /// This function instantiates a \ref DistMap. 
138
    /// \param g is the digraph to which we would like to define the 
139
    /// \ref DistMap.
140
    static DistMap *createDistMap(const GR& g) {
141
      return new DistMap(g);
142
    }
143

	
144
  };
145
  
146
  /// \brief %BellmanFord algorithm class.
147
  ///
148
  /// \ingroup shortest_path
149
  /// This class provides an efficient implementation of the Bellman-Ford 
150
  /// algorithm. The maximum time complexity of the algorithm is
151
  /// <tt>O(ne)</tt>.
152
  ///
153
  /// The Bellman-Ford algorithm solves the single-source shortest path
154
  /// problem when the arcs can have negative lengths, but the digraph
155
  /// should not contain directed cycles with negative total length.
156
  /// If all arc costs are non-negative, consider to use the Dijkstra
157
  /// algorithm instead, since it is more efficient.
158
  ///
159
  /// The arc lengths are passed to the algorithm using a
160
  /// \ref concepts::ReadMap "ReadMap", so it is easy to change it to any 
161
  /// kind of length. The type of the length values is determined by the
162
  /// \ref concepts::ReadMap::Value "Value" type of the length map.
163
  ///
164
  /// There is also a \ref bellmanFord() "function-type interface" for the
165
  /// Bellman-Ford algorithm, which is convenient in the simplier cases and
166
  /// it can be used easier.
167
  ///
168
  /// \tparam GR The type of the digraph the algorithm runs on.
169
  /// The default type is \ref ListDigraph.
170
  /// \tparam LEN A \ref concepts::ReadMap "readable" arc map that specifies
171
  /// the lengths of the arcs. The default map type is
172
  /// \ref concepts::Digraph::ArcMap "GR::ArcMap<int>".
173
#ifdef DOXYGEN
174
  template <typename GR, typename LEN, typename TR>
175
#else
176
  template <typename GR=ListDigraph,
177
            typename LEN=typename GR::template ArcMap<int>,
178
            typename TR=BellmanFordDefaultTraits<GR,LEN> >
179
#endif
180
  class BellmanFord {
181
  public:
182

	
183
    ///The type of the underlying digraph.
184
    typedef typename TR::Digraph Digraph;
185
    
186
    /// \brief The type of the arc lengths.
187
    typedef typename TR::LengthMap::Value Value;
188
    /// \brief The type of the map that stores the arc lengths.
189
    typedef typename TR::LengthMap LengthMap;
190
    /// \brief The type of the map that stores the last
191
    /// arcs of the shortest paths.
192
    typedef typename TR::PredMap PredMap;
193
    /// \brief The type of the map that stores the distances of the nodes.
194
    typedef typename TR::DistMap DistMap;
195
    /// The type of the paths.
196
    typedef PredMapPath<Digraph, PredMap> Path;
197
    ///\brief The \ref BellmanFordDefaultOperationTraits
198
    /// "operation traits class" of the algorithm.
199
    typedef typename TR::OperationTraits OperationTraits;
200

	
201
    ///The \ref BellmanFordDefaultTraits "traits class" of the algorithm.
202
    typedef TR Traits;
203

	
204
  private:
205

	
206
    typedef typename Digraph::Node Node;
207
    typedef typename Digraph::NodeIt NodeIt;
208
    typedef typename Digraph::Arc Arc;
209
    typedef typename Digraph::OutArcIt OutArcIt;
210

	
211
    // Pointer to the underlying digraph.
212
    const Digraph *_gr;
213
    // Pointer to the length map
214
    const LengthMap *_length;
215
    // Pointer to the map of predecessors arcs.
216
    PredMap *_pred;
217
    // Indicates if _pred is locally allocated (true) or not.
218
    bool _local_pred;
219
    // Pointer to the map of distances.
220
    DistMap *_dist;
221
    // Indicates if _dist is locally allocated (true) or not.
222
    bool _local_dist;
223

	
224
    typedef typename Digraph::template NodeMap<bool> MaskMap;
225
    MaskMap *_mask;
226

	
227
    std::vector<Node> _process;
228

	
229
    // Creates the maps if necessary.
230
    void create_maps() {
231
      if(!_pred) {
232
	_local_pred = true;
233
	_pred = Traits::createPredMap(*_gr);
234
      }
235
      if(!_dist) {
236
	_local_dist = true;
237
	_dist = Traits::createDistMap(*_gr);
238
      }
239
      _mask = new MaskMap(*_gr, false);
240
    }
241
    
242
  public :
243
 
244
    typedef BellmanFord Create;
245

	
246
    /// \name Named Template Parameters
247

	
248
    ///@{
249

	
250
    template <class T>
251
    struct SetPredMapTraits : public Traits {
252
      typedef T PredMap;
253
      static PredMap *createPredMap(const Digraph&) {
254
        LEMON_ASSERT(false, "PredMap is not initialized");
255
        return 0; // ignore warnings
256
      }
257
    };
258

	
259
    /// \brief \ref named-templ-param "Named parameter" for setting
260
    /// \c PredMap type.
261
    ///
262
    /// \ref named-templ-param "Named parameter" for setting
263
    /// \c PredMap type.
264
    /// It must conform to the \ref concepts::WriteMap "WriteMap" concept.
265
    template <class T>
266
    struct SetPredMap 
267
      : public BellmanFord< Digraph, LengthMap, SetPredMapTraits<T> > {
268
      typedef BellmanFord< Digraph, LengthMap, SetPredMapTraits<T> > Create;
269
    };
270
    
271
    template <class T>
272
    struct SetDistMapTraits : public Traits {
273
      typedef T DistMap;
274
      static DistMap *createDistMap(const Digraph&) {
275
        LEMON_ASSERT(false, "DistMap is not initialized");
276
        return 0; // ignore warnings
277
      }
278
    };
279

	
280
    /// \brief \ref named-templ-param "Named parameter" for setting
281
    /// \c DistMap type.
282
    ///
283
    /// \ref named-templ-param "Named parameter" for setting
284
    /// \c DistMap type.
285
    /// It must conform to the \ref concepts::WriteMap "WriteMap" concept.
286
    template <class T>
287
    struct SetDistMap 
288
      : public BellmanFord< Digraph, LengthMap, SetDistMapTraits<T> > {
289
      typedef BellmanFord< Digraph, LengthMap, SetDistMapTraits<T> > Create;
290
    };
291

	
292
    template <class T>
293
    struct SetOperationTraitsTraits : public Traits {
294
      typedef T OperationTraits;
295
    };
296
    
297
    /// \brief \ref named-templ-param "Named parameter" for setting 
298
    /// \c OperationTraits type.
299
    ///
300
    /// \ref named-templ-param "Named parameter" for setting
301
    /// \c OperationTraits type.
302
    /// For more information see \ref BellmanFordDefaultOperationTraits.
303
    template <class T>
304
    struct SetOperationTraits
305
      : public BellmanFord< Digraph, LengthMap, SetOperationTraitsTraits<T> > {
306
      typedef BellmanFord< Digraph, LengthMap, SetOperationTraitsTraits<T> >
307
      Create;
308
    };
309
    
310
    ///@}
311

	
312
  protected:
313
    
314
    BellmanFord() {}
315

	
316
  public:      
317
    
318
    /// \brief Constructor.
319
    ///
320
    /// Constructor.
321
    /// \param g The digraph the algorithm runs on.
322
    /// \param length The length map used by the algorithm.
323
    BellmanFord(const Digraph& g, const LengthMap& length) :
324
      _gr(&g), _length(&length),
325
      _pred(0), _local_pred(false),
326
      _dist(0), _local_dist(false), _mask(0) {}
327
    
328
    ///Destructor.
329
    ~BellmanFord() {
330
      if(_local_pred) delete _pred;
331
      if(_local_dist) delete _dist;
332
      if(_mask) delete _mask;
333
    }
334

	
335
    /// \brief Sets the length map.
336
    ///
337
    /// Sets the length map.
338
    /// \return <tt>(*this)</tt>
339
    BellmanFord &lengthMap(const LengthMap &map) {
340
      _length = &map;
341
      return *this;
342
    }
343

	
344
    /// \brief Sets the map that stores the predecessor arcs.
345
    ///
346
    /// Sets the map that stores the predecessor arcs.
347
    /// If you don't use this function before calling \ref run()
348
    /// or \ref init(), an instance will be allocated automatically.
349
    /// The destructor deallocates this automatically allocated map,
350
    /// of course.
351
    /// \return <tt>(*this)</tt>
352
    BellmanFord &predMap(PredMap &map) {
353
      if(_local_pred) {
354
	delete _pred;
355
	_local_pred=false;
356
      }
357
      _pred = &map;
358
      return *this;
359
    }
360

	
361
    /// \brief Sets the map that stores the distances of the nodes.
362
    ///
363
    /// Sets the map that stores the distances of the nodes calculated
364
    /// by the algorithm.
365
    /// If you don't use this function before calling \ref run()
366
    /// or \ref init(), an instance will be allocated automatically.
367
    /// The destructor deallocates this automatically allocated map,
368
    /// of course.
369
    /// \return <tt>(*this)</tt>
370
    BellmanFord &distMap(DistMap &map) {
371
      if(_local_dist) {
372
	delete _dist;
373
	_local_dist=false;
374
      }
375
      _dist = &map;
376
      return *this;
377
    }
378

	
379
    /// \name Execution Control
380
    /// The simplest way to execute the Bellman-Ford algorithm is to use
381
    /// one of the member functions called \ref run().\n
382
    /// If you need better control on the execution, you have to call
383
    /// \ref init() first, then you can add several source nodes
384
    /// with \ref addSource(). Finally the actual path computation can be
385
    /// performed with \ref start(), \ref checkedStart() or
386
    /// \ref limitedStart().
387

	
388
    ///@{
389

	
390
    /// \brief Initializes the internal data structures.
391
    /// 
392
    /// Initializes the internal data structures. The optional parameter
393
    /// is the initial distance of each node.
394
    void init(const Value value = OperationTraits::infinity()) {
395
      create_maps();
396
      for (NodeIt it(*_gr); it != INVALID; ++it) {
397
	_pred->set(it, INVALID);
398
	_dist->set(it, value);
399
      }
400
      _process.clear();
401
      if (OperationTraits::less(value, OperationTraits::infinity())) {
402
	for (NodeIt it(*_gr); it != INVALID; ++it) {
403
	  _process.push_back(it);
404
	  _mask->set(it, true);
405
	}
406
      }
407
    }
408
    
409
    /// \brief Adds a new source node.
410
    ///
411
    /// This function adds a new source node. The optional second parameter
412
    /// is the initial distance of the node.
413
    void addSource(Node source, Value dst = OperationTraits::zero()) {
414
      _dist->set(source, dst);
415
      if (!(*_mask)[source]) {
416
	_process.push_back(source);
417
	_mask->set(source, true);
418
      }
419
    }
420

	
421
    /// \brief Executes one round from the Bellman-Ford algorithm.
422
    ///
423
    /// If the algoritm calculated the distances in the previous round
424
    /// exactly for the paths of at most \c k arcs, then this function
425
    /// will calculate the distances exactly for the paths of at most
426
    /// <tt>k+1</tt> arcs. Performing \c k iterations using this function
427
    /// calculates the shortest path distances exactly for the paths
428
    /// consisting of at most \c k arcs.
429
    ///
430
    /// \warning The paths with limited arc number cannot be retrieved
431
    /// easily with \ref path() or \ref predArc() functions. If you also
432
    /// need the shortest paths and not only the distances, you should
433
    /// store the \ref predMap() "predecessor map" after each iteration
434
    /// and build the path manually.
435
    ///
436
    /// \return \c true when the algorithm have not found more shorter
437
    /// paths.
438
    ///
439
    /// \see ActiveIt
440
    bool processNextRound() {
441
      for (int i = 0; i < int(_process.size()); ++i) {
442
	_mask->set(_process[i], false);
443
      }
444
      std::vector<Node> nextProcess;
445
      std::vector<Value> values(_process.size());
446
      for (int i = 0; i < int(_process.size()); ++i) {
447
	values[i] = (*_dist)[_process[i]];
448
      }
449
      for (int i = 0; i < int(_process.size()); ++i) {
450
	for (OutArcIt it(*_gr, _process[i]); it != INVALID; ++it) {
451
	  Node target = _gr->target(it);
452
	  Value relaxed = OperationTraits::plus(values[i], (*_length)[it]);
453
	  if (OperationTraits::less(relaxed, (*_dist)[target])) {
454
	    _pred->set(target, it);
455
	    _dist->set(target, relaxed);
456
	    if (!(*_mask)[target]) {
457
	      _mask->set(target, true);
458
	      nextProcess.push_back(target);
459
	    }
460
	  }	  
461
	}
462
      }
463
      _process.swap(nextProcess);
464
      return _process.empty();
465
    }
466

	
467
    /// \brief Executes one weak round from the Bellman-Ford algorithm.
468
    ///
469
    /// If the algorithm calculated the distances in the previous round
470
    /// at least for the paths of at most \c k arcs, then this function
471
    /// will calculate the distances at least for the paths of at most
472
    /// <tt>k+1</tt> arcs.
473
    /// This function does not make it possible to calculate the shortest
474
    /// path distances exactly for paths consisting of at most \c k arcs,
475
    /// this is why it is called weak round.
476
    ///
477
    /// \return \c true when the algorithm have not found more shorter
478
    /// paths.
479
    ///
480
    /// \see ActiveIt
481
    bool processNextWeakRound() {
482
      for (int i = 0; i < int(_process.size()); ++i) {
483
	_mask->set(_process[i], false);
484
      }
485
      std::vector<Node> nextProcess;
486
      for (int i = 0; i < int(_process.size()); ++i) {
487
	for (OutArcIt it(*_gr, _process[i]); it != INVALID; ++it) {
488
	  Node target = _gr->target(it);
489
	  Value relaxed = 
490
	    OperationTraits::plus((*_dist)[_process[i]], (*_length)[it]);
491
	  if (OperationTraits::less(relaxed, (*_dist)[target])) {
492
	    _pred->set(target, it);
493
	    _dist->set(target, relaxed);
494
	    if (!(*_mask)[target]) {
495
	      _mask->set(target, true);
496
	      nextProcess.push_back(target);
497
	    }
498
	  }	  
499
	}
500
      }
501
      _process.swap(nextProcess);
502
      return _process.empty();
503
    }
504

	
505
    /// \brief Executes the algorithm.
506
    ///
507
    /// Executes the algorithm.
508
    ///
509
    /// This method runs the Bellman-Ford algorithm from the root node(s)
510
    /// in order to compute the shortest path to each node.
511
    ///
512
    /// The algorithm computes
513
    /// - the shortest path tree (forest),
514
    /// - the distance of each node from the root(s).
515
    ///
516
    /// \pre init() must be called and at least one root node should be
517
    /// added with addSource() before using this function.
518
    void start() {
519
      int num = countNodes(*_gr) - 1;
520
      for (int i = 0; i < num; ++i) {
521
	if (processNextWeakRound()) break;
522
      }
523
    }
524

	
525
    /// \brief Executes the algorithm and checks the negative cycles.
526
    ///
527
    /// Executes the algorithm and checks the negative cycles.
528
    ///
529
    /// This method runs the Bellman-Ford algorithm from the root node(s)
530
    /// in order to compute the shortest path to each node and also checks
531
    /// if the digraph contains cycles with negative total length.
532
    ///
533
    /// The algorithm computes 
534
    /// - the shortest path tree (forest),
535
    /// - the distance of each node from the root(s).
536
    /// 
537
    /// \return \c false if there is a negative cycle in the digraph.
538
    ///
539
    /// \pre init() must be called and at least one root node should be
540
    /// added with addSource() before using this function. 
541
    bool checkedStart() {
542
      int num = countNodes(*_gr);
543
      for (int i = 0; i < num; ++i) {
544
	if (processNextWeakRound()) return true;
545
      }
546
      return _process.empty();
547
    }
548

	
549
    /// \brief Executes the algorithm with arc number limit.
550
    ///
551
    /// Executes the algorithm with arc number limit.
552
    ///
553
    /// This method runs the Bellman-Ford algorithm from the root node(s)
554
    /// in order to compute the shortest path distance for each node
555
    /// using only the paths consisting of at most \c num arcs.
556
    ///
557
    /// The algorithm computes
558
    /// - the limited distance of each node from the root(s),
559
    /// - the predecessor arc for each node.
560
    ///
561
    /// \warning The paths with limited arc number cannot be retrieved
562
    /// easily with \ref path() or \ref predArc() functions. If you also
563
    /// need the shortest paths and not only the distances, you should
564
    /// store the \ref predMap() "predecessor map" after each iteration
565
    /// and build the path manually.
566
    ///
567
    /// \pre init() must be called and at least one root node should be
568
    /// added with addSource() before using this function. 
569
    void limitedStart(int num) {
570
      for (int i = 0; i < num; ++i) {
571
	if (processNextRound()) break;
572
      }
573
    }
574
    
575
    /// \brief Runs the algorithm from the given root node.
576
    ///    
577
    /// This method runs the Bellman-Ford algorithm from the given root
578
    /// node \c s in order to compute the shortest path to each node.
579
    ///
580
    /// The algorithm computes
581
    /// - the shortest path tree (forest),
582
    /// - the distance of each node from the root(s).
583
    ///
584
    /// \note bf.run(s) is just a shortcut of the following code.
585
    /// \code
586
    ///   bf.init();
587
    ///   bf.addSource(s);
588
    ///   bf.start();
589
    /// \endcode
590
    void run(Node s) {
591
      init();
592
      addSource(s);
593
      start();
594
    }
595
    
596
    /// \brief Runs the algorithm from the given root node with arc
597
    /// number limit.
598
    ///    
599
    /// This method runs the Bellman-Ford algorithm from the given root
600
    /// node \c s in order to compute the shortest path distance for each
601
    /// node using only the paths consisting of at most \c num arcs.
602
    ///
603
    /// The algorithm computes
604
    /// - the limited distance of each node from the root(s),
605
    /// - the predecessor arc for each node.
606
    ///
607
    /// \warning The paths with limited arc number cannot be retrieved
608
    /// easily with \ref path() or \ref predArc() functions. If you also
609
    /// need the shortest paths and not only the distances, you should
610
    /// store the \ref predMap() "predecessor map" after each iteration
611
    /// and build the path manually.
612
    ///
613
    /// \note bf.run(s, num) is just a shortcut of the following code.
614
    /// \code
615
    ///   bf.init();
616
    ///   bf.addSource(s);
617
    ///   bf.limitedStart(num);
618
    /// \endcode
619
    void run(Node s, int num) {
620
      init();
621
      addSource(s);
622
      limitedStart(num);
623
    }
624
    
625
    ///@}
626

	
627
    /// \brief LEMON iterator for getting the active nodes.
628
    ///
629
    /// This class provides a common style LEMON iterator that traverses
630
    /// the active nodes of the Bellman-Ford algorithm after the last
631
    /// phase. These nodes should be checked in the next phase to
632
    /// find augmenting arcs outgoing from them.
633
    class ActiveIt {
634
    public:
635

	
636
      /// \brief Constructor.
637
      ///
638
      /// Constructor for getting the active nodes of the given BellmanFord
639
      /// instance. 
640
      ActiveIt(const BellmanFord& algorithm) : _algorithm(&algorithm)
641
      {
642
        _index = _algorithm->_process.size() - 1;
643
      }
644

	
645
      /// \brief Invalid constructor.
646
      ///
647
      /// Invalid constructor.
648
      ActiveIt(Invalid) : _algorithm(0), _index(-1) {}
649

	
650
      /// \brief Conversion to \c Node.
651
      ///
652
      /// Conversion to \c Node.
653
      operator Node() const { 
654
        return _index >= 0 ? _algorithm->_process[_index] : INVALID;
655
      }
656

	
657
      /// \brief Increment operator.
658
      ///
659
      /// Increment operator.
660
      ActiveIt& operator++() {
661
        --_index;
662
        return *this; 
663
      }
664

	
665
      bool operator==(const ActiveIt& it) const { 
666
        return static_cast<Node>(*this) == static_cast<Node>(it); 
667
      }
668
      bool operator!=(const ActiveIt& it) const { 
669
        return static_cast<Node>(*this) != static_cast<Node>(it); 
670
      }
671
      bool operator<(const ActiveIt& it) const { 
672
        return static_cast<Node>(*this) < static_cast<Node>(it); 
673
      }
674
      
675
    private:
676
      const BellmanFord* _algorithm;
677
      int _index;
678
    };
679
    
680
    /// \name Query Functions
681
    /// The result of the Bellman-Ford algorithm can be obtained using these
682
    /// functions.\n
683
    /// Either \ref run() or \ref init() should be called before using them.
684
    
685
    ///@{
686

	
687
    /// \brief The shortest path to the given node.
688
    ///    
689
    /// Gives back the shortest path to the given node from the root(s).
690
    ///
691
    /// \warning \c t should be reached from the root(s).
692
    ///
693
    /// \pre Either \ref run() or \ref init() must be called before
694
    /// using this function.
695
    Path path(Node t) const
696
    {
697
      return Path(*_gr, *_pred, t);
698
    }
699
	  
700
    /// \brief The distance of the given node from the root(s).
701
    ///
702
    /// Returns the distance of the given node from the root(s).
703
    ///
704
    /// \warning If node \c v is not reached from the root(s), then
705
    /// the return value of this function is undefined.
706
    ///
707
    /// \pre Either \ref run() or \ref init() must be called before
708
    /// using this function.
709
    Value dist(Node v) const { return (*_dist)[v]; }
710

	
711
    /// \brief Returns the 'previous arc' of the shortest path tree for
712
    /// the given node.
713
    ///
714
    /// This function returns the 'previous arc' of the shortest path
715
    /// tree for node \c v, i.e. it returns the last arc of a
716
    /// shortest path from a root to \c v. It is \c INVALID if \c v
717
    /// is not reached from the root(s) or if \c v is a root.
718
    ///
719
    /// The shortest path tree used here is equal to the shortest path
720
    /// tree used in \ref predNode() and \predMap().
721
    ///
722
    /// \pre Either \ref run() or \ref init() must be called before
723
    /// using this function.
724
    Arc predArc(Node v) const { return (*_pred)[v]; }
725

	
726
    /// \brief Returns the 'previous node' of the shortest path tree for
727
    /// the given node.
728
    ///
729
    /// This function returns the 'previous node' of the shortest path
730
    /// tree for node \c v, i.e. it returns the last but one node of
731
    /// a shortest path from a root to \c v. It is \c INVALID if \c v
732
    /// is not reached from the root(s) or if \c v is a root.
733
    ///
734
    /// The shortest path tree used here is equal to the shortest path
735
    /// tree used in \ref predArc() and \predMap().
736
    ///
737
    /// \pre Either \ref run() or \ref init() must be called before
738
    /// using this function.
739
    Node predNode(Node v) const { 
740
      return (*_pred)[v] == INVALID ? INVALID : _gr->source((*_pred)[v]); 
741
    }
742
    
743
    /// \brief Returns a const reference to the node map that stores the
744
    /// distances of the nodes.
745
    ///
746
    /// Returns a const reference to the node map that stores the distances
747
    /// of the nodes calculated by the algorithm.
748
    ///
749
    /// \pre Either \ref run() or \ref init() must be called before
750
    /// using this function.
751
    const DistMap &distMap() const { return *_dist;}
752
 
753
    /// \brief Returns a const reference to the node map that stores the
754
    /// predecessor arcs.
755
    ///
756
    /// Returns a const reference to the node map that stores the predecessor
757
    /// arcs, which form the shortest path tree (forest).
758
    ///
759
    /// \pre Either \ref run() or \ref init() must be called before
760
    /// using this function.
761
    const PredMap &predMap() const { return *_pred; }
762
 
763
    /// \brief Checks if a node is reached from the root(s).
764
    ///
765
    /// Returns \c true if \c v is reached from the root(s).
766
    ///
767
    /// \pre Either \ref run() or \ref init() must be called before
768
    /// using this function.
769
    bool reached(Node v) const {
770
      return (*_dist)[v] != OperationTraits::infinity();
771
    }
772

	
773
    /// \brief Gives back a negative cycle.
774
    ///    
775
    /// This function gives back a directed cycle with negative total
776
    /// length if the algorithm has already found one.
777
    /// Otherwise it gives back an empty path.
778
    lemon::Path<Digraph> negativeCycle() {
779
      typename Digraph::template NodeMap<int> state(*_gr, -1);
780
      lemon::Path<Digraph> cycle;
781
      for (int i = 0; i < int(_process.size()); ++i) {
782
        if (state[_process[i]] != -1) continue;
783
        for (Node v = _process[i]; (*_pred)[v] != INVALID;
784
             v = _gr->source((*_pred)[v])) {
785
          if (state[v] == i) {
786
            cycle.addFront((*_pred)[v]);
787
            for (Node u = _gr->source((*_pred)[v]); u != v;
788
                 u = _gr->source((*_pred)[u])) {
789
              cycle.addFront((*_pred)[u]);
790
            }
791
            return cycle;
792
          }
793
          else if (state[v] >= 0) {
794
            break;
795
          }
796
          state[v] = i;
797
        }
798
      }
799
      return cycle;
800
    }
801
    
802
    ///@}
803
  };
804
 
805
  /// \brief Default traits class of bellmanFord() function.
806
  ///
807
  /// Default traits class of bellmanFord() function.
808
  /// \tparam GR The type of the digraph.
809
  /// \tparam LEN The type of the length map.
810
  template <typename GR, typename LEN>
811
  struct BellmanFordWizardDefaultTraits {
812
    /// The type of the digraph the algorithm runs on. 
813
    typedef GR Digraph;
814

	
815
    /// \brief The type of the map that stores the arc lengths.
816
    ///
817
    /// The type of the map that stores the arc lengths.
818
    /// It must meet the \ref concepts::ReadMap "ReadMap" concept.
819
    typedef LEN LengthMap;
820

	
821
    /// The type of the arc lengths.
822
    typedef typename LEN::Value Value;
823

	
824
    /// \brief Operation traits for Bellman-Ford algorithm.
825
    ///
826
    /// It defines the used operations and the infinity value for the
827
    /// given \c Value type.
828
    /// \see BellmanFordDefaultOperationTraits
829
    typedef BellmanFordDefaultOperationTraits<Value> OperationTraits;
830

	
831
    /// \brief The type of the map that stores the last
832
    /// arcs of the shortest paths.
833
    /// 
834
    /// The type of the map that stores the last arcs of the shortest paths.
835
    /// It must conform to the \ref concepts::WriteMap "WriteMap" concept.
836
    typedef typename GR::template NodeMap<typename GR::Arc> PredMap;
837

	
838
    /// \brief Instantiates a \c PredMap.
839
    /// 
840
    /// This function instantiates a \ref PredMap.
841
    /// \param g is the digraph to which we would like to define the
842
    /// \ref PredMap.
843
    static PredMap *createPredMap(const GR &g) {
844
      return new PredMap(g);
845
    }
846

	
847
    /// \brief The type of the map that stores the distances of the nodes.
848
    ///
849
    /// The type of the map that stores the distances of the nodes.
850
    /// It must conform to the \ref concepts::WriteMap "WriteMap" concept.
851
    typedef typename GR::template NodeMap<Value> DistMap;
852

	
853
    /// \brief Instantiates a \c DistMap.
854
    ///
855
    /// This function instantiates a \ref DistMap. 
856
    /// \param g is the digraph to which we would like to define the
857
    /// \ref DistMap.
858
    static DistMap *createDistMap(const GR &g) {
859
      return new DistMap(g);
860
    }
861

	
862
    ///The type of the shortest paths.
863

	
864
    ///The type of the shortest paths.
865
    ///It must meet the \ref concepts::Path "Path" concept.
866
    typedef lemon::Path<Digraph> Path;
867
  };
868
  
869
  /// \brief Default traits class used by BellmanFordWizard.
870
  ///
871
  /// Default traits class used by BellmanFordWizard.
872
  /// \tparam GR The type of the digraph.
873
  /// \tparam LEN The type of the length map.
874
  template <typename GR, typename LEN>
875
  class BellmanFordWizardBase 
876
    : public BellmanFordWizardDefaultTraits<GR, LEN> {
877

	
878
    typedef BellmanFordWizardDefaultTraits<GR, LEN> Base;
879
  protected:
880
    // Type of the nodes in the digraph.
881
    typedef typename Base::Digraph::Node Node;
882

	
883
    // Pointer to the underlying digraph.
884
    void *_graph;
885
    // Pointer to the length map
886
    void *_length;
887
    // Pointer to the map of predecessors arcs.
888
    void *_pred;
889
    // Pointer to the map of distances.
890
    void *_dist;
891
    //Pointer to the shortest path to the target node.
892
    void *_path;
893
    //Pointer to the distance of the target node.
894
    void *_di;
895

	
896
    public:
897
    /// Constructor.
898
    
899
    /// This constructor does not require parameters, it initiates
900
    /// all of the attributes to default values \c 0.
901
    BellmanFordWizardBase() :
902
      _graph(0), _length(0), _pred(0), _dist(0), _path(0), _di(0) {}
903

	
904
    /// Constructor.
905
    
906
    /// This constructor requires two parameters,
907
    /// others are initiated to \c 0.
908
    /// \param gr The digraph the algorithm runs on.
909
    /// \param len The length map.
910
    BellmanFordWizardBase(const GR& gr, 
911
			  const LEN& len) :
912
      _graph(reinterpret_cast<void*>(const_cast<GR*>(&gr))), 
913
      _length(reinterpret_cast<void*>(const_cast<LEN*>(&len))), 
914
      _pred(0), _dist(0), _path(0), _di(0) {}
915

	
916
  };
917
  
918
  /// \brief Auxiliary class for the function-type interface of the
919
  /// \ref BellmanFord "Bellman-Ford" algorithm.
920
  ///
921
  /// This auxiliary class is created to implement the
922
  /// \ref bellmanFord() "function-type interface" of the
923
  /// \ref BellmanFord "Bellman-Ford" algorithm.
924
  /// It does not have own \ref run() method, it uses the
925
  /// functions and features of the plain \ref BellmanFord.
926
  ///
927
  /// This class should only be used through the \ref bellmanFord()
928
  /// function, which makes it easier to use the algorithm.
929
  template<class TR>
930
  class BellmanFordWizard : public TR {
931
    typedef TR Base;
932

	
933
    typedef typename TR::Digraph Digraph;
934

	
935
    typedef typename Digraph::Node Node;
936
    typedef typename Digraph::NodeIt NodeIt;
937
    typedef typename Digraph::Arc Arc;
938
    typedef typename Digraph::OutArcIt ArcIt;
939
    
940
    typedef typename TR::LengthMap LengthMap;
941
    typedef typename LengthMap::Value Value;
942
    typedef typename TR::PredMap PredMap;
943
    typedef typename TR::DistMap DistMap;
944
    typedef typename TR::Path Path;
945

	
946
  public:
947
    /// Constructor.
948
    BellmanFordWizard() : TR() {}
949

	
950
    /// \brief Constructor that requires parameters.
951
    ///
952
    /// Constructor that requires parameters.
953
    /// These parameters will be the default values for the traits class.
954
    /// \param gr The digraph the algorithm runs on.
955
    /// \param len The length map.
956
    BellmanFordWizard(const Digraph& gr, const LengthMap& len) 
957
      : TR(gr, len) {}
958

	
959
    /// \brief Copy constructor
960
    BellmanFordWizard(const TR &b) : TR(b) {}
961

	
962
    ~BellmanFordWizard() {}
963

	
964
    /// \brief Runs the Bellman-Ford algorithm from the given source node.
965
    ///    
966
    /// This method runs the Bellman-Ford algorithm from the given source
967
    /// node in order to compute the shortest path to each node.
968
    void run(Node s) {
969
      BellmanFord<Digraph,LengthMap,TR> 
970
	bf(*reinterpret_cast<const Digraph*>(Base::_graph), 
971
           *reinterpret_cast<const LengthMap*>(Base::_length));
972
      if (Base::_pred) bf.predMap(*reinterpret_cast<PredMap*>(Base::_pred));
973
      if (Base::_dist) bf.distMap(*reinterpret_cast<DistMap*>(Base::_dist));
974
      bf.run(s);
975
    }
976

	
977
    /// \brief Runs the Bellman-Ford algorithm to find the shortest path
978
    /// between \c s and \c t.
979
    ///
980
    /// This method runs the Bellman-Ford algorithm from node \c s
981
    /// in order to compute the shortest path to node \c t.
982
    /// Actually, it computes the shortest path to each node, but using
983
    /// this function you can retrieve the distance and the shortest path
984
    /// for a single target node easier.
985
    ///
986
    /// \return \c true if \c t is reachable form \c s.
987
    bool run(Node s, Node t) {
988
      BellmanFord<Digraph,LengthMap,TR>
989
        bf(*reinterpret_cast<const Digraph*>(Base::_graph),
990
           *reinterpret_cast<const LengthMap*>(Base::_length));
991
      if (Base::_pred) bf.predMap(*reinterpret_cast<PredMap*>(Base::_pred));
992
      if (Base::_dist) bf.distMap(*reinterpret_cast<DistMap*>(Base::_dist));
993
      bf.run(s);
994
      if (Base::_path) *reinterpret_cast<Path*>(Base::_path) = bf.path(t);
995
      if (Base::_di) *reinterpret_cast<Value*>(Base::_di) = bf.dist(t);
996
      return bf.reached(t);
997
    }
998

	
999
    template<class T>
1000
    struct SetPredMapBase : public Base {
1001
      typedef T PredMap;
1002
      static PredMap *createPredMap(const Digraph &) { return 0; };
1003
      SetPredMapBase(const TR &b) : TR(b) {}
1004
    };
1005
    
1006
    /// \brief \ref named-templ-param "Named parameter" for setting
1007
    /// the predecessor map.
1008
    ///
1009
    /// \ref named-templ-param "Named parameter" for setting
1010
    /// the map that stores the predecessor arcs of the nodes.
1011
    template<class T>
1012
    BellmanFordWizard<SetPredMapBase<T> > predMap(const T &t) {
1013
      Base::_pred=reinterpret_cast<void*>(const_cast<T*>(&t));
1014
      return BellmanFordWizard<SetPredMapBase<T> >(*this);
1015
    }
1016
    
1017
    template<class T>
1018
    struct SetDistMapBase : public Base {
1019
      typedef T DistMap;
1020
      static DistMap *createDistMap(const Digraph &) { return 0; };
1021
      SetDistMapBase(const TR &b) : TR(b) {}
1022
    };
1023
    
1024
    /// \brief \ref named-templ-param "Named parameter" for setting
1025
    /// the distance map.
1026
    ///
1027
    /// \ref named-templ-param "Named parameter" for setting
1028
    /// the map that stores the distances of the nodes calculated
1029
    /// by the algorithm.
1030
    template<class T>
1031
    BellmanFordWizard<SetDistMapBase<T> > distMap(const T &t) {
1032
      Base::_dist=reinterpret_cast<void*>(const_cast<T*>(&t));
1033
      return BellmanFordWizard<SetDistMapBase<T> >(*this);
1034
    }
1035

	
1036
    template<class T>
1037
    struct SetPathBase : public Base {
1038
      typedef T Path;
1039
      SetPathBase(const TR &b) : TR(b) {}
1040
    };
1041

	
1042
    /// \brief \ref named-func-param "Named parameter" for getting
1043
    /// the shortest path to the target node.
1044
    ///
1045
    /// \ref named-func-param "Named parameter" for getting
1046
    /// the shortest path to the target node.
1047
    template<class T>
1048
    BellmanFordWizard<SetPathBase<T> > path(const T &t)
1049
    {
1050
      Base::_path=reinterpret_cast<void*>(const_cast<T*>(&t));
1051
      return BellmanFordWizard<SetPathBase<T> >(*this);
1052
    }
1053

	
1054
    /// \brief \ref named-func-param "Named parameter" for getting
1055
    /// the distance of the target node.
1056
    ///
1057
    /// \ref named-func-param "Named parameter" for getting
1058
    /// the distance of the target node.
1059
    BellmanFordWizard dist(const Value &d)
1060
    {
1061
      Base::_di=reinterpret_cast<void*>(const_cast<Value*>(&d));
1062
      return *this;
1063
    }
1064
    
1065
  };
1066
  
1067
  /// \brief Function type interface for the \ref BellmanFord "Bellman-Ford"
1068
  /// algorithm.
1069
  ///
1070
  /// \ingroup shortest_path
1071
  /// Function type interface for the \ref BellmanFord "Bellman-Ford"
1072
  /// algorithm.
1073
  ///
1074
  /// This function also has several \ref named-templ-func-param 
1075
  /// "named parameters", they are declared as the members of class 
1076
  /// \ref BellmanFordWizard.
1077
  /// The following examples show how to use these parameters.
1078
  /// \code
1079
  ///   // Compute shortest path from node s to each node
1080
  ///   bellmanFord(g,length).predMap(preds).distMap(dists).run(s);
1081
  ///
1082
  ///   // Compute shortest path from s to t
1083
  ///   bool reached = bellmanFord(g,length).path(p).dist(d).run(s,t);
1084
  /// \endcode
1085
  /// \warning Don't forget to put the \ref BellmanFordWizard::run() "run()"
1086
  /// to the end of the parameter list.
1087
  /// \sa BellmanFordWizard
1088
  /// \sa BellmanFord
1089
  template<typename GR, typename LEN>
1090
  BellmanFordWizard<BellmanFordWizardBase<GR,LEN> >
1091
  bellmanFord(const GR& digraph,
1092
	      const LEN& length)
1093
  {
1094
    return BellmanFordWizard<BellmanFordWizardBase<GR,LEN> >(digraph, length);
1095
  }
1096

	
1097
} //END OF NAMESPACE LEMON
1098

	
1099
#endif
1100

	
Ignore white space 6 line context
1
/* -*- mode: C++; indent-tabs-mode: nil; -*-
2
 *
3
 * This file is a part of LEMON, a generic C++ optimization library.
4
 *
5
 * Copyright (C) 2003-2009
6
 * Egervary Jeno Kombinatorikus Optimalizalasi Kutatocsoport
7
 * (Egervary Research Group on Combinatorial Optimization, EGRES).
8
 *
9
 * Permission to use, modify and distribute this software is granted
10
 * provided that this copyright notice appears in all copies. For
11
 * precise terms see the accompanying LICENSE file.
12
 *
13
 * This software is provided "AS IS" with no warranty of any kind,
14
 * express or implied, and with no claim as to its suitability for any
15
 * purpose.
16
 *
17
 */
18

	
19
#ifndef LEMON_BINOM_HEAP_H
20
#define LEMON_BINOM_HEAP_H
21

	
22
///\file
23
///\ingroup heaps
24
///\brief Binomial Heap implementation.
25

	
26
#include <vector>
27
#include <utility>
28
#include <functional>
29
#include <lemon/math.h>
30
#include <lemon/counter.h>
31

	
32
namespace lemon {
33

	
34
  /// \ingroup heaps
35
  ///
36
  ///\brief Binomial heap data structure.
37
  ///
38
  /// This class implements the \e binomial \e heap data structure.
39
  /// It fully conforms to the \ref concepts::Heap "heap concept".
40
  ///
41
  /// The methods \ref increase() and \ref erase() are not efficient
42
  /// in a binomial heap. In case of many calls of these operations,
43
  /// it is better to use other heap structure, e.g. \ref BinHeap
44
  /// "binary heap".
45
  ///
46
  /// \tparam PR Type of the priorities of the items.
47
  /// \tparam IM A read-writable item map with \c int values, used
48
  /// internally to handle the cross references.
49
  /// \tparam CMP A functor class for comparing the priorities.
50
  /// The default is \c std::less<PR>.
51
#ifdef DOXYGEN
52
  template <typename PR, typename IM, typename CMP>
53
#else
54
  template <typename PR, typename IM, typename CMP = std::less<PR> >
55
#endif
56
  class BinomHeap {
57
  public:
58
    /// Type of the item-int map.
59
    typedef IM ItemIntMap;
60
    /// Type of the priorities.
61
    typedef PR Prio;
62
    /// Type of the items stored in the heap.
63
    typedef typename ItemIntMap::Key Item;
64
    /// Functor type for comparing the priorities.
65
    typedef CMP Compare;
66

	
67
    /// \brief Type to represent the states of the items.
68
    ///
69
    /// Each item has a state associated to it. It can be "in heap",
70
    /// "pre-heap" or "post-heap". The latter two are indifferent from the
71
    /// heap's point of view, but may be useful to the user.
72
    ///
73
    /// The item-int map must be initialized in such way that it assigns
74
    /// \c PRE_HEAP (<tt>-1</tt>) to any element to be put in the heap.
75
    enum State {
76
      IN_HEAP = 0,    ///< = 0.
77
      PRE_HEAP = -1,  ///< = -1.
78
      POST_HEAP = -2  ///< = -2.
79
    };
80

	
81
  private:
82
    class Store;
83

	
84
    std::vector<Store> _data;
85
    int _min, _head;
86
    ItemIntMap &_iim;
87
    Compare _comp;
88
    int _num_items;
89

	
90
  public:
91
    /// \brief Constructor.
92
    ///
93
    /// Constructor.
94
    /// \param map A map that assigns \c int values to the items.
95
    /// It is used internally to handle the cross references.
96
    /// The assigned value must be \c PRE_HEAP (<tt>-1</tt>) for each item.
97
    explicit BinomHeap(ItemIntMap &map)
98
      : _min(0), _head(-1), _iim(map), _num_items(0) {}
99

	
100
    /// \brief Constructor.
101
    ///
102
    /// Constructor.
103
    /// \param map A map that assigns \c int values to the items.
104
    /// It is used internally to handle the cross references.
105
    /// The assigned value must be \c PRE_HEAP (<tt>-1</tt>) for each item.
106
    /// \param comp The function object used for comparing the priorities.
107
    BinomHeap(ItemIntMap &map, const Compare &comp)
108
      : _min(0), _head(-1), _iim(map), _comp(comp), _num_items(0) {}
109

	
110
    /// \brief The number of items stored in the heap.
111
    ///
112
    /// This function returns the number of items stored in the heap.
113
    int size() const { return _num_items; }
114

	
115
    /// \brief Check if the heap is empty.
116
    ///
117
    /// This function returns \c true if the heap is empty.
118
    bool empty() const { return _num_items==0; }
119

	
120
    /// \brief Make the heap empty.
121
    ///
122
    /// This functon makes the heap empty.
123
    /// It does not change the cross reference map. If you want to reuse
124
    /// a heap that is not surely empty, you should first clear it and
125
    /// then you should set the cross reference map to \c PRE_HEAP
126
    /// for each item.
127
    void clear() {
128
      _data.clear(); _min=0; _num_items=0; _head=-1;
129
    }
130

	
131
    /// \brief Set the priority of an item or insert it, if it is
132
    /// not stored in the heap.
133
    ///
134
    /// This method sets the priority of the given item if it is
135
    /// already stored in the heap. Otherwise it inserts the given
136
    /// item into the heap with the given priority.
137
    /// \param item The item.
138
    /// \param value The priority.
139
    void set (const Item& item, const Prio& value) {
140
      int i=_iim[item];
141
      if ( i >= 0 && _data[i].in ) {
142
        if ( _comp(value, _data[i].prio) ) decrease(item, value);
143
        if ( _comp(_data[i].prio, value) ) increase(item, value);
144
      } else push(item, value);
145
    }
146

	
147
    /// \brief Insert an item into the heap with the given priority.
148
    ///
149
    /// This function inserts the given item into the heap with the
150
    /// given priority.
151
    /// \param item The item to insert.
152
    /// \param value The priority of the item.
153
    /// \pre \e item must not be stored in the heap.
154
    void push (const Item& item, const Prio& value) {
155
      int i=_iim[item];
156
      if ( i<0 ) {
157
        int s=_data.size();
158
        _iim.set( item,s );
159
        Store st;
160
        st.name=item;
161
        st.prio=value;
162
        _data.push_back(st);
163
        i=s;
164
      }
165
      else {
166
        _data[i].parent=_data[i].right_neighbor=_data[i].child=-1;
167
        _data[i].degree=0;
168
        _data[i].in=true;
169
        _data[i].prio=value;
170
      }
171

	
172
      if( 0==_num_items ) {
173
        _head=i;
174
        _min=i;
175
      } else {
176
        merge(i);
177
        if( _comp(_data[i].prio, _data[_min].prio) ) _min=i;
178
      }
179
      ++_num_items;
180
    }
181

	
182
    /// \brief Return the item having minimum priority.
183
    ///
184
    /// This function returns the item having minimum priority.
185
    /// \pre The heap must be non-empty.
186
    Item top() const { return _data[_min].name; }
187

	
188
    /// \brief The minimum priority.
189
    ///
190
    /// This function returns the minimum priority.
191
    /// \pre The heap must be non-empty.
192
    Prio prio() const { return _data[_min].prio; }
193

	
194
    /// \brief The priority of the given item.
195
    ///
196
    /// This function returns the priority of the given item.
197
    /// \param item The item.
198
    /// \pre \e item must be in the heap.
199
    const Prio& operator[](const Item& item) const {
200
      return _data[_iim[item]].prio;
201
    }
202

	
203
    /// \brief Remove the item having minimum priority.
204
    ///
205
    /// This function removes the item having minimum priority.
206
    /// \pre The heap must be non-empty.
207
    void pop() {
208
      _data[_min].in=false;
209

	
210
      int head_child=-1;
211
      if ( _data[_min].child!=-1 ) {
212
        int child=_data[_min].child;
213
        int neighb;
214
        while( child!=-1 ) {
215
          neighb=_data[child].right_neighbor;
216
          _data[child].parent=-1;
217
          _data[child].right_neighbor=head_child;
218
          head_child=child;
219
          child=neighb;
220
        }
221
      }
222

	
223
      if ( _data[_head].right_neighbor==-1 ) {
224
        // there was only one root
225
        _head=head_child;
226
      }
227
      else {
228
        // there were more roots
229
        if( _head!=_min )  { unlace(_min); }
230
        else { _head=_data[_head].right_neighbor; }
231
        merge(head_child);
232
      }
233
      _min=findMin();
234
      --_num_items;
235
    }
236

	
237
    /// \brief Remove the given item from the heap.
238
    ///
239
    /// This function removes the given item from the heap if it is
240
    /// already stored.
241
    /// \param item The item to delete.
242
    /// \pre \e item must be in the heap.
243
    void erase (const Item& item) {
244
      int i=_iim[item];
245
      if ( i >= 0 && _data[i].in ) {
246
        decrease( item, _data[_min].prio-1 );
247
        pop();
248
      }
249
    }
250

	
251
    /// \brief Decrease the priority of an item to the given value.
252
    ///
253
    /// This function decreases the priority of an item to the given value.
254
    /// \param item The item.
255
    /// \param value The priority.
256
    /// \pre \e item must be stored in the heap with priority at least \e value.
257
    void decrease (Item item, const Prio& value) {
258
      int i=_iim[item];
259
      int p=_data[i].parent;
260
      _data[i].prio=value;
261
      
262
      while( p!=-1 && _comp(value, _data[p].prio) ) {
263
        _data[i].name=_data[p].name;
264
        _data[i].prio=_data[p].prio;
265
        _data[p].name=item;
266
        _data[p].prio=value;
267
        _iim[_data[i].name]=i;
268
        i=p;
269
        p=_data[p].parent;
270
      }
271
      _iim[item]=i;
272
      if ( _comp(value, _data[_min].prio) ) _min=i;
273
    }
274

	
275
    /// \brief Increase the priority of an item to the given value.
276
    ///
277
    /// This function increases the priority of an item to the given value.
278
    /// \param item The item.
279
    /// \param value The priority.
280
    /// \pre \e item must be stored in the heap with priority at most \e value.
281
    void increase (Item item, const Prio& value) {
282
      erase(item);
283
      push(item, value);
284
    }
285

	
286
    /// \brief Return the state of an item.
287
    ///
288
    /// This method returns \c PRE_HEAP if the given item has never
289
    /// been in the heap, \c IN_HEAP if it is in the heap at the moment,
290
    /// and \c POST_HEAP otherwise.
291
    /// In the latter case it is possible that the item will get back
292
    /// to the heap again.
293
    /// \param item The item.
294
    State state(const Item &item) const {
295
      int i=_iim[item];
296
      if( i>=0 ) {
297
        if ( _data[i].in ) i=0;
298
        else i=-2;
299
      }
300
      return State(i);
301
    }
302

	
303
    /// \brief Set the state of an item in the heap.
304
    ///
305
    /// This function sets the state of the given item in the heap.
306
    /// It can be used to manually clear the heap when it is important
307
    /// to achive better time complexity.
308
    /// \param i The item.
309
    /// \param st The state. It should not be \c IN_HEAP.
310
    void state(const Item& i, State st) {
311
      switch (st) {
312
      case POST_HEAP:
313
      case PRE_HEAP:
314
        if (state(i) == IN_HEAP) {
315
          erase(i);
316
        }
317
        _iim[i] = st;
318
        break;
319
      case IN_HEAP:
320
        break;
321
      }
322
    }
323

	
324
  private:
325
    
326
    // Find the minimum of the roots
327
    int findMin() {
328
      if( _head!=-1 ) {
329
        int min_loc=_head, min_val=_data[_head].prio;
330
        for( int x=_data[_head].right_neighbor; x!=-1;
331
             x=_data[x].right_neighbor ) {
332
          if( _comp( _data[x].prio,min_val ) ) {
333
            min_val=_data[x].prio;
334
            min_loc=x;
335
          }
336
        }
337
        return min_loc;
338
      }
339
      else return -1;
340
    }
341

	
342
    // Merge the heap with another heap starting at the given position
343
    void merge(int a) {
344
      if( _head==-1 || a==-1 ) return;
345
      if( _data[a].right_neighbor==-1 &&
346
          _data[a].degree<=_data[_head].degree ) {
347
        _data[a].right_neighbor=_head;
348
        _head=a;
349
      } else {
350
        interleave(a);
351
      }
352
      if( _data[_head].right_neighbor==-1 ) return;
353
      
354
      int x=_head;
355
      int x_prev=-1, x_next=_data[x].right_neighbor;
356
      while( x_next!=-1 ) {
357
        if( _data[x].degree!=_data[x_next].degree ||
358
            ( _data[x_next].right_neighbor!=-1 &&
359
              _data[_data[x_next].right_neighbor].degree==_data[x].degree ) ) {
360
          x_prev=x;
361
          x=x_next;
362
        }
363
        else {
364
          if( _comp(_data[x_next].prio,_data[x].prio) ) {
365
            if( x_prev==-1 ) {
366
              _head=x_next;
367
            } else {
368
              _data[x_prev].right_neighbor=x_next;
369
            }
370
            fuse(x,x_next);
371
            x=x_next;
372
          }
373
          else {
374
            _data[x].right_neighbor=_data[x_next].right_neighbor;
375
            fuse(x_next,x);
376
          }
377
        }
378
        x_next=_data[x].right_neighbor;
379
      }
380
    }
381

	
382
    // Interleave the elements of the given list into the list of the roots
383
    void interleave(int a) {
384
      int p=_head, q=a;
385
      int curr=_data.size();
386
      _data.push_back(Store());
387
      
388
      while( p!=-1 || q!=-1 ) {
389
        if( q==-1 || ( p!=-1 && _data[p].degree<_data[q].degree ) ) {
390
          _data[curr].right_neighbor=p;
391
          curr=p;
392
          p=_data[p].right_neighbor;
393
        }
394
        else {
395
          _data[curr].right_neighbor=q;
396
          curr=q;
397
          q=_data[q].right_neighbor;
398
        }
399
      }
400
      
401
      _head=_data.back().right_neighbor;
402
      _data.pop_back();
403
    }
404

	
405
    // Lace node a under node b
406
    void fuse(int a, int b) {
407
      _data[a].parent=b;
408
      _data[a].right_neighbor=_data[b].child;
409
      _data[b].child=a;
410

	
411
      ++_data[b].degree;
412
    }
413

	
414
    // Unlace node a (if it has siblings)
415
    void unlace(int a) {
416
      int neighb=_data[a].right_neighbor;
417
      int other=_head;
418

	
419
      while( _data[other].right_neighbor!=a )
420
        other=_data[other].right_neighbor;
421
      _data[other].right_neighbor=neighb;
422
    }
423

	
424
  private:
425

	
426
    class Store {
427
      friend class BinomHeap;
428

	
429
      Item name;
430
      int parent;
431
      int right_neighbor;
432
      int child;
433
      int degree;
434
      bool in;
435
      Prio prio;
436

	
437
      Store() : parent(-1), right_neighbor(-1), child(-1), degree(0),
438
        in(true) {}
439
    };
440
  };
441

	
442
} //namespace lemon
443

	
444
#endif //LEMON_BINOM_HEAP_H
445

	
Ignore white space 6 line context
1
/* -*- mode: C++; indent-tabs-mode: nil; -*-
2
 *
3
 * This file is a part of LEMON, a generic C++ optimization library.
4
 *
5
 * Copyright (C) 2003-2009
6
 * Egervary Jeno Kombinatorikus Optimalizalasi Kutatocsoport
7
 * (Egervary Research Group on Combinatorial Optimization, EGRES).
8
 *
9
 * Permission to use, modify and distribute this software is granted
10
 * provided that this copyright notice appears in all copies. For
11
 * precise terms see the accompanying LICENSE file.
12
 *
13
 * This software is provided "AS IS" with no warranty of any kind,
14
 * express or implied, and with no claim as to its suitability for any
15
 * purpose.
16
 *
17
 */
18

	
19
#ifndef LEMON_BUCKET_HEAP_H
20
#define LEMON_BUCKET_HEAP_H
21

	
22
///\ingroup heaps
23
///\file
24
///\brief Bucket heap implementation.
25

	
26
#include <vector>
27
#include <utility>
28
#include <functional>
29

	
30
namespace lemon {
31

	
32
  namespace _bucket_heap_bits {
33

	
34
    template <bool MIN>
35
    struct DirectionTraits {
36
      static bool less(int left, int right) {
37
        return left < right;
38
      }
39
      static void increase(int& value) {
40
        ++value;
41
      }
42
    };
43

	
44
    template <>
45
    struct DirectionTraits<false> {
46
      static bool less(int left, int right) {
47
        return left > right;
48
      }
49
      static void increase(int& value) {
50
        --value;
51
      }
52
    };
53

	
54
  }
55

	
56
  /// \ingroup heaps
57
  ///
58
  /// \brief Bucket heap data structure.
59
  ///
60
  /// This class implements the \e bucket \e heap data structure.
61
  /// It practically conforms to the \ref concepts::Heap "heap concept",
62
  /// but it has some limitations.
63
  ///
64
  /// The bucket heap is a very simple structure. It can store only
65
  /// \c int priorities and it maintains a list of items for each priority
66
  /// in the range <tt>[0..C)</tt>. So it should only be used when the
67
  /// priorities are small. It is not intended to use as a Dijkstra heap.
68
  ///
69
  /// \tparam IM A read-writable item map with \c int values, used
70
  /// internally to handle the cross references.
71
  /// \tparam MIN Indicate if the heap is a \e min-heap or a \e max-heap.
72
  /// The default is \e min-heap. If this parameter is set to \c false,
73
  /// then the comparison is reversed, so the top(), prio() and pop()
74
  /// functions deal with the item having maximum priority instead of the
75
  /// minimum.
76
  ///
77
  /// \sa SimpleBucketHeap
78
  template <typename IM, bool MIN = true>
79
  class BucketHeap {
80

	
81
  public:
82

	
83
    /// Type of the item-int map.
84
    typedef IM ItemIntMap;
85
    /// Type of the priorities.
86
    typedef int Prio;
87
    /// Type of the items stored in the heap.
88
    typedef typename ItemIntMap::Key Item;
89
    /// Type of the item-priority pairs.
90
    typedef std::pair<Item,Prio> Pair;
91

	
92
  private:
93

	
94
    typedef _bucket_heap_bits::DirectionTraits<MIN> Direction;
95

	
96
  public:
97

	
98
    /// \brief Type to represent the states of the items.
99
    ///
100
    /// Each item has a state associated to it. It can be "in heap",
101
    /// "pre-heap" or "post-heap". The latter two are indifferent from the
102
    /// heap's point of view, but may be useful to the user.
103
    ///
104
    /// The item-int map must be initialized in such way that it assigns
105
    /// \c PRE_HEAP (<tt>-1</tt>) to any element to be put in the heap.
106
    enum State {
107
      IN_HEAP = 0,    ///< = 0.
108
      PRE_HEAP = -1,  ///< = -1.
109
      POST_HEAP = -2  ///< = -2.
110
    };
111

	
112
  public:
113

	
114
    /// \brief Constructor.
115
    ///
116
    /// Constructor.
117
    /// \param map A map that assigns \c int values to the items.
118
    /// It is used internally to handle the cross references.
119
    /// The assigned value must be \c PRE_HEAP (<tt>-1</tt>) for each item.
120
    explicit BucketHeap(ItemIntMap &map) : _iim(map), _minimum(0) {}
121

	
122
    /// \brief The number of items stored in the heap.
123
    ///
124
    /// This function returns the number of items stored in the heap.
125
    int size() const { return _data.size(); }
126

	
127
    /// \brief Check if the heap is empty.
128
    ///
129
    /// This function returns \c true if the heap is empty.
130
    bool empty() const { return _data.empty(); }
131

	
132
    /// \brief Make the heap empty.
133
    ///
134
    /// This functon makes the heap empty.
135
    /// It does not change the cross reference map. If you want to reuse
136
    /// a heap that is not surely empty, you should first clear it and
137
    /// then you should set the cross reference map to \c PRE_HEAP
138
    /// for each item.
139
    void clear() {
140
      _data.clear(); _first.clear(); _minimum = 0;
141
    }
142

	
143
  private:
144

	
145
    void relocateLast(int idx) {
146
      if (idx + 1 < int(_data.size())) {
147
        _data[idx] = _data.back();
148
        if (_data[idx].prev != -1) {
149
          _data[_data[idx].prev].next = idx;
150
        } else {
151
          _first[_data[idx].value] = idx;
152
        }
153
        if (_data[idx].next != -1) {
154
          _data[_data[idx].next].prev = idx;
155
        }
156
        _iim[_data[idx].item] = idx;
157
      }
158
      _data.pop_back();
159
    }
160

	
161
    void unlace(int idx) {
162
      if (_data[idx].prev != -1) {
163
        _data[_data[idx].prev].next = _data[idx].next;
164
      } else {
165
        _first[_data[idx].value] = _data[idx].next;
166
      }
167
      if (_data[idx].next != -1) {
168
        _data[_data[idx].next].prev = _data[idx].prev;
169
      }
170
    }
171

	
172
    void lace(int idx) {
173
      if (int(_first.size()) <= _data[idx].value) {
174
        _first.resize(_data[idx].value + 1, -1);
175
      }
176
      _data[idx].next = _first[_data[idx].value];
177
      if (_data[idx].next != -1) {
178
        _data[_data[idx].next].prev = idx;
179
      }
180
      _first[_data[idx].value] = idx;
181
      _data[idx].prev = -1;
182
    }
183

	
184
  public:
185

	
186
    /// \brief Insert a pair of item and priority into the heap.
187
    ///
188
    /// This function inserts \c p.first to the heap with priority
189
    /// \c p.second.
190
    /// \param p The pair to insert.
191
    /// \pre \c p.first must not be stored in the heap.
192
    void push(const Pair& p) {
193
      push(p.first, p.second);
194
    }
195

	
196
    /// \brief Insert an item into the heap with the given priority.
197
    ///
198
    /// This function inserts the given item into the heap with the
199
    /// given priority.
200
    /// \param i The item to insert.
201
    /// \param p The priority of the item.
202
    /// \pre \e i must not be stored in the heap.
203
    void push(const Item &i, const Prio &p) {
204
      int idx = _data.size();
205
      _iim[i] = idx;
206
      _data.push_back(BucketItem(i, p));
207
      lace(idx);
208
      if (Direction::less(p, _minimum)) {
209
        _minimum = p;
210
      }
211
    }
212

	
213
    /// \brief Return the item having minimum priority.
214
    ///
215
    /// This function returns the item having minimum priority.
216
    /// \pre The heap must be non-empty.
217
    Item top() const {
218
      while (_first[_minimum] == -1) {
219
        Direction::increase(_minimum);
220
      }
221
      return _data[_first[_minimum]].item;
222
    }
223

	
224
    /// \brief The minimum priority.
225
    ///
226
    /// This function returns the minimum priority.
227
    /// \pre The heap must be non-empty.
228
    Prio prio() const {
229
      while (_first[_minimum] == -1) {
230
        Direction::increase(_minimum);
231
      }
232
      return _minimum;
233
    }
234

	
235
    /// \brief Remove the item having minimum priority.
236
    ///
237
    /// This function removes the item having minimum priority.
238
    /// \pre The heap must be non-empty.
239
    void pop() {
240
      while (_first[_minimum] == -1) {
241
        Direction::increase(_minimum);
242
      }
243
      int idx = _first[_minimum];
244
      _iim[_data[idx].item] = -2;
245
      unlace(idx);
246
      relocateLast(idx);
247
    }
248

	
249
    /// \brief Remove the given item from the heap.
250
    ///
251
    /// This function removes the given item from the heap if it is
252
    /// already stored.
253
    /// \param i The item to delete.
254
    /// \pre \e i must be in the heap.
255
    void erase(const Item &i) {
256
      int idx = _iim[i];
257
      _iim[_data[idx].item] = -2;
258
      unlace(idx);
259
      relocateLast(idx);
260
    }
261

	
262
    /// \brief The priority of the given item.
263
    ///
264
    /// This function returns the priority of the given item.
265
    /// \param i The item.
266
    /// \pre \e i must be in the heap.
267
    Prio operator[](const Item &i) const {
268
      int idx = _iim[i];
269
      return _data[idx].value;
270
    }
271

	
272
    /// \brief Set the priority of an item or insert it, if it is
273
    /// not stored in the heap.
274
    ///
275
    /// This method sets the priority of the given item if it is
276
    /// already stored in the heap. Otherwise it inserts the given
277
    /// item into the heap with the given priority.
278
    /// \param i The item.
279
    /// \param p The priority.
280
    void set(const Item &i, const Prio &p) {
281
      int idx = _iim[i];
282
      if (idx < 0) {
283
        push(i, p);
284
      } else if (Direction::less(p, _data[idx].value)) {
285
        decrease(i, p);
286
      } else {
287
        increase(i, p);
288
      }
289
    }
290

	
291
    /// \brief Decrease the priority of an item to the given value.
292
    ///
293
    /// This function decreases the priority of an item to the given value.
294
    /// \param i The item.
295
    /// \param p The priority.
296
    /// \pre \e i must be stored in the heap with priority at least \e p.
297
    void decrease(const Item &i, const Prio &p) {
298
      int idx = _iim[i];
299
      unlace(idx);
300
      _data[idx].value = p;
301
      if (Direction::less(p, _minimum)) {
302
        _minimum = p;
303
      }
304
      lace(idx);
305
    }
306

	
307
    /// \brief Increase the priority of an item to the given value.
308
    ///
309
    /// This function increases the priority of an item to the given value.
310
    /// \param i The item.
311
    /// \param p The priority.
312
    /// \pre \e i must be stored in the heap with priority at most \e p.
313
    void increase(const Item &i, const Prio &p) {
314
      int idx = _iim[i];
315
      unlace(idx);
316
      _data[idx].value = p;
317
      lace(idx);
318
    }
319

	
320
    /// \brief Return the state of an item.
321
    ///
322
    /// This method returns \c PRE_HEAP if the given item has never
323
    /// been in the heap, \c IN_HEAP if it is in the heap at the moment,
324
    /// and \c POST_HEAP otherwise.
325
    /// In the latter case it is possible that the item will get back
326
    /// to the heap again.
327
    /// \param i The item.
328
    State state(const Item &i) const {
329
      int idx = _iim[i];
330
      if (idx >= 0) idx = 0;
331
      return State(idx);
332
    }
333

	
334
    /// \brief Set the state of an item in the heap.
335
    ///
336
    /// This function sets the state of the given item in the heap.
337
    /// It can be used to manually clear the heap when it is important
338
    /// to achive better time complexity.
339
    /// \param i The item.
340
    /// \param st The state. It should not be \c IN_HEAP.
341
    void state(const Item& i, State st) {
342
      switch (st) {
343
      case POST_HEAP:
344
      case PRE_HEAP:
345
        if (state(i) == IN_HEAP) {
346
          erase(i);
347
        }
348
        _iim[i] = st;
349
        break;
350
      case IN_HEAP:
351
        break;
352
      }
353
    }
354

	
355
  private:
356

	
357
    struct BucketItem {
358
      BucketItem(const Item& _item, int _value)
359
        : item(_item), value(_value) {}
360

	
361
      Item item;
362
      int value;
363

	
364
      int prev, next;
365
    };
366

	
367
    ItemIntMap& _iim;
368
    std::vector<int> _first;
369
    std::vector<BucketItem> _data;
370
    mutable int _minimum;
371

	
372
  }; // class BucketHeap
373

	
374
  /// \ingroup heaps
375
  ///
376
  /// \brief Simplified bucket heap data structure.
377
  ///
378
  /// This class implements a simplified \e bucket \e heap data
379
  /// structure. It does not provide some functionality, but it is
380
  /// faster and simpler than BucketHeap. The main difference is
381
  /// that BucketHeap stores a doubly-linked list for each key while
382
  /// this class stores only simply-linked lists. It supports erasing
383
  /// only for the item having minimum priority and it does not support
384
  /// key increasing and decreasing.
385
  ///
386
  /// Note that this implementation does not conform to the
387
  /// \ref concepts::Heap "heap concept" due to the lack of some 
388
  /// functionality.
389
  ///
390
  /// \tparam IM A read-writable item map with \c int values, used
391
  /// internally to handle the cross references.
392
  /// \tparam MIN Indicate if the heap is a \e min-heap or a \e max-heap.
393
  /// The default is \e min-heap. If this parameter is set to \c false,
394
  /// then the comparison is reversed, so the top(), prio() and pop()
395
  /// functions deal with the item having maximum priority instead of the
396
  /// minimum.
397
  ///
398
  /// \sa BucketHeap
399
  template <typename IM, bool MIN = true >
400
  class SimpleBucketHeap {
401

	
402
  public:
403

	
404
    /// Type of the item-int map.
405
    typedef IM ItemIntMap;
406
    /// Type of the priorities.
407
    typedef int Prio;
408
    /// Type of the items stored in the heap.
409
    typedef typename ItemIntMap::Key Item;
410
    /// Type of the item-priority pairs.
411
    typedef std::pair<Item,Prio> Pair;
412

	
413
  private:
414

	
415
    typedef _bucket_heap_bits::DirectionTraits<MIN> Direction;
416

	
417
  public:
418

	
419
    /// \brief Type to represent the states of the items.
420
    ///
421
    /// Each item has a state associated to it. It can be "in heap",
422
    /// "pre-heap" or "post-heap". The latter two are indifferent from the
423
    /// heap's point of view, but may be useful to the user.
424
    ///
425
    /// The item-int map must be initialized in such way that it assigns
426
    /// \c PRE_HEAP (<tt>-1</tt>) to any element to be put in the heap.
427
    enum State {
428
      IN_HEAP = 0,    ///< = 0.
429
      PRE_HEAP = -1,  ///< = -1.
430
      POST_HEAP = -2  ///< = -2.
431
    };
432

	
433
  public:
434

	
435
    /// \brief Constructor.
436
    ///
437
    /// Constructor.
438
    /// \param map A map that assigns \c int values to the items.
439
    /// It is used internally to handle the cross references.
440
    /// The assigned value must be \c PRE_HEAP (<tt>-1</tt>) for each item.
441
    explicit SimpleBucketHeap(ItemIntMap &map)
442
      : _iim(map), _free(-1), _num(0), _minimum(0) {}
443

	
444
    /// \brief The number of items stored in the heap.
445
    ///
446
    /// This function returns the number of items stored in the heap.
447
    int size() const { return _num; }
448

	
449
    /// \brief Check if the heap is empty.
450
    ///
451
    /// This function returns \c true if the heap is empty.
452
    bool empty() const { return _num == 0; }
453

	
454
    /// \brief Make the heap empty.
455
    ///
456
    /// This functon makes the heap empty.
457
    /// It does not change the cross reference map. If you want to reuse
458
    /// a heap that is not surely empty, you should first clear it and
459
    /// then you should set the cross reference map to \c PRE_HEAP
460
    /// for each item.
461
    void clear() {
462
      _data.clear(); _first.clear(); _free = -1; _num = 0; _minimum = 0;
463
    }
464

	
465
    /// \brief Insert a pair of item and priority into the heap.
466
    ///
467
    /// This function inserts \c p.first to the heap with priority
468
    /// \c p.second.
469
    /// \param p The pair to insert.
470
    /// \pre \c p.first must not be stored in the heap.
471
    void push(const Pair& p) {
472
      push(p.first, p.second);
473
    }
474

	
475
    /// \brief Insert an item into the heap with the given priority.
476
    ///
477
    /// This function inserts the given item into the heap with the
478
    /// given priority.
479
    /// \param i The item to insert.
480
    /// \param p The priority of the item.
481
    /// \pre \e i must not be stored in the heap.
482
    void push(const Item &i, const Prio &p) {
483
      int idx;
484
      if (_free == -1) {
485
        idx = _data.size();
486
        _data.push_back(BucketItem(i));
487
      } else {
488
        idx = _free;
489
        _free = _data[idx].next;
490
        _data[idx].item = i;
491
      }
492
      _iim[i] = idx;
493
      if (p >= int(_first.size())) _first.resize(p + 1, -1);
494
      _data[idx].next = _first[p];
495
      _first[p] = idx;
496
      if (Direction::less(p, _minimum)) {
497
        _minimum = p;
498
      }
499
      ++_num;
500
    }
501

	
502
    /// \brief Return the item having minimum priority.
503
    ///
504
    /// This function returns the item having minimum priority.
505
    /// \pre The heap must be non-empty.
506
    Item top() const {
507
      while (_first[_minimum] == -1) {
508
        Direction::increase(_minimum);
509
      }
510
      return _data[_first[_minimum]].item;
511
    }
512

	
513
    /// \brief The minimum priority.
514
    ///
515
    /// This function returns the minimum priority.
516
    /// \pre The heap must be non-empty.
517
    Prio prio() const {
518
      while (_first[_minimum] == -1) {
519
        Direction::increase(_minimum);
520
      }
521
      return _minimum;
522
    }
523

	
524
    /// \brief Remove the item having minimum priority.
525
    ///
526
    /// This function removes the item having minimum priority.
527
    /// \pre The heap must be non-empty.
528
    void pop() {
529
      while (_first[_minimum] == -1) {
530
        Direction::increase(_minimum);
531
      }
532
      int idx = _first[_minimum];
533
      _iim[_data[idx].item] = -2;
534
      _first[_minimum] = _data[idx].next;
535
      _data[idx].next = _free;
536
      _free = idx;
537
      --_num;
538
    }
539

	
540
    /// \brief The priority of the given item.
541
    ///
542
    /// This function returns the priority of the given item.
543
    /// \param i The item.
544
    /// \pre \e i must be in the heap.
545
    /// \warning This operator is not a constant time function because
546
    /// it scans the whole data structure to find the proper value.
547
    Prio operator[](const Item &i) const {
548
      for (int k = 0; k < int(_first.size()); ++k) {
549
        int idx = _first[k];
550
        while (idx != -1) {
551
          if (_data[idx].item == i) {
552
            return k;
553
          }
554
          idx = _data[idx].next;
555
        }
556
      }
557
      return -1;
558
    }
559

	
560
    /// \brief Return the state of an item.
561
    ///
562
    /// This method returns \c PRE_HEAP if the given item has never
563
    /// been in the heap, \c IN_HEAP if it is in the heap at the moment,
564
    /// and \c POST_HEAP otherwise.
565
    /// In the latter case it is possible that the item will get back
566
    /// to the heap again.
567
    /// \param i The item.
568
    State state(const Item &i) const {
569
      int idx = _iim[i];
570
      if (idx >= 0) idx = 0;
571
      return State(idx);
572
    }
573

	
574
  private:
575

	
576
    struct BucketItem {
577
      BucketItem(const Item& _item)
578
        : item(_item) {}
579

	
580
      Item item;
581
      int next;
582
    };
583

	
584
    ItemIntMap& _iim;
585
    std::vector<int> _first;
586
    std::vector<BucketItem> _data;
587
    int _free, _num;
588
    mutable int _minimum;
589

	
590
  }; // class SimpleBucketHeap
591

	
592
}
593

	
594
#endif
Ignore white space 192 line context
1
/* -*- mode: C++; indent-tabs-mode: nil; -*-
2
 *
3
 * This file is a part of LEMON, a generic C++ optimization library.
4
 *
5
 * Copyright (C) 2003-2009
6
 * Egervary Jeno Kombinatorikus Optimalizalasi Kutatocsoport
7
 * (Egervary Research Group on Combinatorial Optimization, EGRES).
8
 *
9
 * Permission to use, modify and distribute this software is granted
10
 * provided that this copyright notice appears in all copies. For
11
 * precise terms see the accompanying LICENSE file.
12
 *
13
 * This software is provided "AS IS" with no warranty of any kind,
14
 * express or implied, and with no claim as to its suitability for any
15
 * purpose.
16
 *
17
 */
18

	
19
#ifndef LEMON_FIB_HEAP_H
20
#define LEMON_FIB_HEAP_H
21

	
22
///\file
23
///\ingroup heaps
24
///\brief Fibonacci heap implementation.
25

	
26
#include <vector>
27
#include <utility>
28
#include <functional>
29
#include <lemon/math.h>
30

	
31
namespace lemon {
32

	
33
  /// \ingroup heaps
34
  ///
35
  /// \brief Fibonacci heap data structure.
36
  ///
37
  /// This class implements the \e Fibonacci \e heap data structure.
38
  /// It fully conforms to the \ref concepts::Heap "heap concept".
39
  ///
40
  /// The methods \ref increase() and \ref erase() are not efficient in a
41
  /// Fibonacci heap. In case of many calls of these operations, it is
42
  /// better to use other heap structure, e.g. \ref BinHeap "binary heap".
43
  ///
44
  /// \tparam PR Type of the priorities of the items.
45
  /// \tparam IM A read-writable item map with \c int values, used
46
  /// internally to handle the cross references.
47
  /// \tparam CMP A functor class for comparing the priorities.
48
  /// The default is \c std::less<PR>.
49
#ifdef DOXYGEN
50
  template <typename PR, typename IM, typename CMP>
51
#else
52
  template <typename PR, typename IM, typename CMP = std::less<PR> >
53
#endif
54
  class FibHeap {
55
  public:
56

	
57
    /// Type of the item-int map.
58
    typedef IM ItemIntMap;
59
    /// Type of the priorities.
60
    typedef PR Prio;
61
    /// Type of the items stored in the heap.
62
    typedef typename ItemIntMap::Key Item;
63
    /// Type of the item-priority pairs.
64
    typedef std::pair<Item,Prio> Pair;
65
    /// Functor type for comparing the priorities.
66
    typedef CMP Compare;
67

	
68
  private:
69
    class Store;
70

	
71
    std::vector<Store> _data;
72
    int _minimum;
73
    ItemIntMap &_iim;
74
    Compare _comp;
75
    int _num;
76

	
77
  public:
78

	
79
    /// \brief Type to represent the states of the items.
80
    ///
81
    /// Each item has a state associated to it. It can be "in heap",
82
    /// "pre-heap" or "post-heap". The latter two are indifferent from the
83
    /// heap's point of view, but may be useful to the user.
84
    ///
85
    /// The item-int map must be initialized in such way that it assigns
86
    /// \c PRE_HEAP (<tt>-1</tt>) to any element to be put in the heap.
87
    enum State {
88
      IN_HEAP = 0,    ///< = 0.
89
      PRE_HEAP = -1,  ///< = -1.
90
      POST_HEAP = -2  ///< = -2.
91
    };
92

	
93
    /// \brief Constructor.
94
    ///
95
    /// Constructor.
96
    /// \param map A map that assigns \c int values to the items.
97
    /// It is used internally to handle the cross references.
98
    /// The assigned value must be \c PRE_HEAP (<tt>-1</tt>) for each item.
99
    explicit FibHeap(ItemIntMap &map)
100
      : _minimum(0), _iim(map), _num() {}
101

	
102
    /// \brief Constructor.
103
    ///
104
    /// Constructor.
105
    /// \param map A map that assigns \c int values to the items.
106
    /// It is used internally to handle the cross references.
107
    /// The assigned value must be \c PRE_HEAP (<tt>-1</tt>) for each item.
108
    /// \param comp The function object used for comparing the priorities.
109
    FibHeap(ItemIntMap &map, const Compare &comp)
110
      : _minimum(0), _iim(map), _comp(comp), _num() {}
111

	
112
    /// \brief The number of items stored in the heap.
113
    ///
114
    /// This function returns the number of items stored in the heap.
115
    int size() const { return _num; }
116

	
117
    /// \brief Check if the heap is empty.
118
    ///
119
    /// This function returns \c true if the heap is empty.
120
    bool empty() const { return _num==0; }
121

	
122
    /// \brief Make the heap empty.
123
    ///
124
    /// This functon makes the heap empty.
125
    /// It does not change the cross reference map. If you want to reuse
126
    /// a heap that is not surely empty, you should first clear it and
127
    /// then you should set the cross reference map to \c PRE_HEAP
128
    /// for each item.
129
    void clear() {
130
      _data.clear(); _minimum = 0; _num = 0;
131
    }
132

	
133
    /// \brief Insert an item into the heap with the given priority.
134
    ///
135
    /// This function inserts the given item into the heap with the
136
    /// given priority.
137
    /// \param item The item to insert.
138
    /// \param prio The priority of the item.
139
    /// \pre \e item must not be stored in the heap.
140
    void push (const Item& item, const Prio& prio) {
141
      int i=_iim[item];
142
      if ( i < 0 ) {
143
        int s=_data.size();
144
        _iim.set( item, s );
145
        Store st;
146
        st.name=item;
147
        _data.push_back(st);
148
        i=s;
149
      } else {
150
        _data[i].parent=_data[i].child=-1;
151
        _data[i].degree=0;
152
        _data[i].in=true;
153
        _data[i].marked=false;
154
      }
155

	
156
      if ( _num ) {
157
        _data[_data[_minimum].right_neighbor].left_neighbor=i;
158
        _data[i].right_neighbor=_data[_minimum].right_neighbor;
159
        _data[_minimum].right_neighbor=i;
160
        _data[i].left_neighbor=_minimum;
161
        if ( _comp( prio, _data[_minimum].prio) ) _minimum=i;
162
      } else {
163
        _data[i].right_neighbor=_data[i].left_neighbor=i;
164
        _minimum=i;
165
      }
166
      _data[i].prio=prio;
167
      ++_num;
168
    }
169

	
170
    /// \brief Return the item having minimum priority.
171
    ///
172
    /// This function returns the item having minimum priority.
173
    /// \pre The heap must be non-empty.
174
    Item top() const { return _data[_minimum].name; }
175

	
176
    /// \brief The minimum priority.
177
    ///
178
    /// This function returns the minimum priority.
179
    /// \pre The heap must be non-empty.
180
    Prio prio() const { return _data[_minimum].prio; }
181

	
182
    /// \brief Remove the item having minimum priority.
183
    ///
184
    /// This function removes the item having minimum priority.
185
    /// \pre The heap must be non-empty.
186
    void pop() {
187
      /*The first case is that there are only one root.*/
188
      if ( _data[_minimum].left_neighbor==_minimum ) {
189
        _data[_minimum].in=false;
190
        if ( _data[_minimum].degree!=0 ) {
191
          makeRoot(_data[_minimum].child);
192
          _minimum=_data[_minimum].child;
193
          balance();
194
        }
195
      } else {
196
        int right=_data[_minimum].right_neighbor;
197
        unlace(_minimum);
198
        _data[_minimum].in=false;
199
        if ( _data[_minimum].degree > 0 ) {
200
          int left=_data[_minimum].left_neighbor;
201
          int child=_data[_minimum].child;
202
          int last_child=_data[child].left_neighbor;
203

	
204
          makeRoot(child);
205

	
206
          _data[left].right_neighbor=child;
207
          _data[child].left_neighbor=left;
208
          _data[right].left_neighbor=last_child;
209
          _data[last_child].right_neighbor=right;
210
        }
211
        _minimum=right;
212
        balance();
213
      } // the case where there are more roots
214
      --_num;
215
    }
216

	
217
    /// \brief Remove the given item from the heap.
218
    ///
219
    /// This function removes the given item from the heap if it is
220
    /// already stored.
221
    /// \param item The item to delete.
222
    /// \pre \e item must be in the heap.
223
    void erase (const Item& item) {
224
      int i=_iim[item];
225

	
226
      if ( i >= 0 && _data[i].in ) {
227
        if ( _data[i].parent!=-1 ) {
228
          int p=_data[i].parent;
229
          cut(i,p);
230
          cascade(p);
231
        }
232
        _minimum=i;     //As if its prio would be -infinity
233
        pop();
234
      }
235
    }
236

	
237
    /// \brief The priority of the given item.
238
    ///
239
    /// This function returns the priority of the given item.
240
    /// \param item The item.
241
    /// \pre \e item must be in the heap.
242
    Prio operator[](const Item& item) const {
243
      return _data[_iim[item]].prio;
244
    }
245

	
246
    /// \brief Set the priority of an item or insert it, if it is
247
    /// not stored in the heap.
248
    ///
249
    /// This method sets the priority of the given item if it is
250
    /// already stored in the heap. Otherwise it inserts the given
251
    /// item into the heap with the given priority.
252
    /// \param item The item.
253
    /// \param prio The priority.
254
    void set (const Item& item, const Prio& prio) {
255
      int i=_iim[item];
256
      if ( i >= 0 && _data[i].in ) {
257
        if ( _comp(prio, _data[i].prio) ) decrease(item, prio);
258
        if ( _comp(_data[i].prio, prio) ) increase(item, prio);
259
      } else push(item, prio);
260
    }
261

	
262
    /// \brief Decrease the priority of an item to the given value.
263
    ///
264
    /// This function decreases the priority of an item to the given value.
265
    /// \param item The item.
266
    /// \param prio The priority.
267
    /// \pre \e item must be stored in the heap with priority at least \e prio.
268
    void decrease (const Item& item, const Prio& prio) {
269
      int i=_iim[item];
270
      _data[i].prio=prio;
271
      int p=_data[i].parent;
272

	
273
      if ( p!=-1 && _comp(prio, _data[p].prio) ) {
274
        cut(i,p);
275
        cascade(p);
276
      }
277
      if ( _comp(prio, _data[_minimum].prio) ) _minimum=i;
278
    }
279

	
280
    /// \brief Increase the priority of an item to the given value.
281
    ///
282
    /// This function increases the priority of an item to the given value.
283
    /// \param item The item.
284
    /// \param prio The priority.
285
    /// \pre \e item must be stored in the heap with priority at most \e prio.
286
    void increase (const Item& item, const Prio& prio) {
287
      erase(item);
288
      push(item, prio);
289
    }
290

	
291
    /// \brief Return the state of an item.
292
    ///
293
    /// This method returns \c PRE_HEAP if the given item has never
294
    /// been in the heap, \c IN_HEAP if it is in the heap at the moment,
295
    /// and \c POST_HEAP otherwise.
296
    /// In the latter case it is possible that the item will get back
297
    /// to the heap again.
298
    /// \param item The item.
299
    State state(const Item &item) const {
300
      int i=_iim[item];
301
      if( i>=0 ) {
302
        if ( _data[i].in ) i=0;
303
        else i=-2;
304
      }
305
      return State(i);
306
    }
307

	
308
    /// \brief Set the state of an item in the heap.
309
    ///
310
    /// This function sets the state of the given item in the heap.
311
    /// It can be used to manually clear the heap when it is important
312
    /// to achive better time complexity.
313
    /// \param i The item.
314
    /// \param st The state. It should not be \c IN_HEAP.
315
    void state(const Item& i, State st) {
316
      switch (st) {
317
      case POST_HEAP:
318
      case PRE_HEAP:
319
        if (state(i) == IN_HEAP) {
320
          erase(i);
321
        }
322
        _iim[i] = st;
323
        break;
324
      case IN_HEAP:
325
        break;
326
      }
327
    }
328

	
329
  private:
330

	
331
    void balance() {
332

	
333
      int maxdeg=int( std::floor( 2.08*log(double(_data.size()))))+1;
334

	
335
      std::vector<int> A(maxdeg,-1);
336

	
337
      /*
338
       *Recall that now minimum does not point to the minimum prio element.
339
       *We set minimum to this during balance().
340
       */
341
      int anchor=_data[_minimum].left_neighbor;
342
      int next=_minimum;
343
      bool end=false;
344

	
345
      do {
346
        int active=next;
347
        if ( anchor==active ) end=true;
348
        int d=_data[active].degree;
349
        next=_data[active].right_neighbor;
350

	
351
        while (A[d]!=-1) {
352
          if( _comp(_data[active].prio, _data[A[d]].prio) ) {
353
            fuse(active,A[d]);
354
          } else {
355
            fuse(A[d],active);
356
            active=A[d];
357
          }
358
          A[d]=-1;
359
          ++d;
360
        }
361
        A[d]=active;
362
      } while ( !end );
363

	
364

	
365
      while ( _data[_minimum].parent >=0 )
366
        _minimum=_data[_minimum].parent;
367
      int s=_minimum;
368
      int m=_minimum;
369
      do {
370
        if ( _comp(_data[s].prio, _data[_minimum].prio) ) _minimum=s;
371
        s=_data[s].right_neighbor;
372
      } while ( s != m );
373
    }
374

	
375
    void makeRoot(int c) {
376
      int s=c;
377
      do {
378
        _data[s].parent=-1;
379
        s=_data[s].right_neighbor;
380
      } while ( s != c );
381
    }
382

	
383
    void cut(int a, int b) {
384
      /*
385
       *Replacing a from the children of b.
386
       */
387
      --_data[b].degree;
388

	
389
      if ( _data[b].degree !=0 ) {
390
        int child=_data[b].child;
391
        if ( child==a )
392
          _data[b].child=_data[child].right_neighbor;
393
        unlace(a);
394
      }
395

	
396

	
397
      /*Lacing a to the roots.*/
398
      int right=_data[_minimum].right_neighbor;
399
      _data[_minimum].right_neighbor=a;
400
      _data[a].left_neighbor=_minimum;
401
      _data[a].right_neighbor=right;
402
      _data[right].left_neighbor=a;
403

	
404
      _data[a].parent=-1;
405
      _data[a].marked=false;
406
    }
407

	
408
    void cascade(int a) {
409
      if ( _data[a].parent!=-1 ) {
410
        int p=_data[a].parent;
411

	
412
        if ( _data[a].marked==false ) _data[a].marked=true;
413
        else {
414
          cut(a,p);
415
          cascade(p);
416
        }
417
      }
418
    }
419

	
420
    void fuse(int a, int b) {
421
      unlace(b);
422

	
423
      /*Lacing b under a.*/
424
      _data[b].parent=a;
425

	
426
      if (_data[a].degree==0) {
427
        _data[b].left_neighbor=b;
428
        _data[b].right_neighbor=b;
429
        _data[a].child=b;
430
      } else {
431
        int child=_data[a].child;
432
        int last_child=_data[child].left_neighbor;
433
        _data[child].left_neighbor=b;
434
        _data[b].right_neighbor=child;
435
        _data[last_child].right_neighbor=b;
436
        _data[b].left_neighbor=last_child;
437
      }
438

	
439
      ++_data[a].degree;
440

	
441
      _data[b].marked=false;
442
    }
443

	
444
    /*
445
     *It is invoked only if a has siblings.
446
     */
447
    void unlace(int a) {
448
      int leftn=_data[a].left_neighbor;
449
      int rightn=_data[a].right_neighbor;
450
      _data[leftn].right_neighbor=rightn;
451
      _data[rightn].left_neighbor=leftn;
452
    }
453

	
454

	
455
    class Store {
456
      friend class FibHeap;
457

	
458
      Item name;
459
      int parent;
460
      int left_neighbor;
461
      int right_neighbor;
462
      int child;
463
      int degree;
464
      bool marked;
465
      bool in;
466
      Prio prio;
467

	
468
      Store() : parent(-1), child(-1), degree(), marked(false), in(true) {}
469
    };
470
  };
471

	
472
} //namespace lemon
473

	
474
#endif //LEMON_FIB_HEAP_H
475

	
Ignore white space 6 line context
1
/* -*- mode: C++; indent-tabs-mode: nil; -*-
2
 *
3
 * This file is a part of LEMON, a generic C++ optimization library.
4
 *
5
 * Copyright (C) 2003-2009
6
 * Egervary Jeno Kombinatorikus Optimalizalasi Kutatocsoport
7
 * (Egervary Research Group on Combinatorial Optimization, EGRES).
8
 *
9
 * Permission to use, modify and distribute this software is granted
10
 * provided that this copyright notice appears in all copies. For
11
 * precise terms see the accompanying LICENSE file.
12
 *
13
 * This software is provided "AS IS" with no warranty of any kind,
14
 * express or implied, and with no claim as to its suitability for any
15
 * purpose.
16
 *
17
 */
18

	
19
#ifndef LEMON_FOURARY_HEAP_H
20
#define LEMON_FOURARY_HEAP_H
21

	
22
///\ingroup heaps
23
///\file
24
///\brief Fourary heap implementation.
25

	
26
#include <vector>
27
#include <utility>
28
#include <functional>
29

	
30
namespace lemon {
31

	
32
  /// \ingroup heaps
33
  ///
34
  ///\brief Fourary heap data structure.
35
  ///
36
  /// This class implements the \e fourary \e heap data structure.
37
  /// It fully conforms to the \ref concepts::Heap "heap concept".
38
  ///
39
  /// The fourary heap is a specialization of the \ref KaryHeap "K-ary heap"
40
  /// for <tt>K=4</tt>. It is similar to the \ref BinHeap "binary heap",
41
  /// but its nodes have at most four children, instead of two.
42
  ///
43
  /// \tparam PR Type of the priorities of the items.
44
  /// \tparam IM A read-writable item map with \c int values, used
45
  /// internally to handle the cross references.
46
  /// \tparam CMP A functor class for comparing the priorities.
47
  /// The default is \c std::less<PR>.
48
  ///
49
  ///\sa BinHeap
50
  ///\sa KaryHeap
51
#ifdef DOXYGEN
52
  template <typename PR, typename IM, typename CMP>
53
#else
54
  template <typename PR, typename IM, typename CMP = std::less<PR> >
55
#endif
56
  class FouraryHeap {
57
  public:
58
    /// Type of the item-int map.
59
    typedef IM ItemIntMap;
60
    /// Type of the priorities.
61
    typedef PR Prio;
62
    /// Type of the items stored in the heap.
63
    typedef typename ItemIntMap::Key Item;
64
    /// Type of the item-priority pairs.
65
    typedef std::pair<Item,Prio> Pair;
66
    /// Functor type for comparing the priorities.
67
    typedef CMP Compare;
68

	
69
    /// \brief Type to represent the states of the items.
70
    ///
71
    /// Each item has a state associated to it. It can be "in heap",
72
    /// "pre-heap" or "post-heap". The latter two are indifferent from the
73
    /// heap's point of view, but may be useful to the user.
74
    ///
75
    /// The item-int map must be initialized in such way that it assigns
76
    /// \c PRE_HEAP (<tt>-1</tt>) to any element to be put in the heap.
77
    enum State {
78
      IN_HEAP = 0,    ///< = 0.
79
      PRE_HEAP = -1,  ///< = -1.
80
      POST_HEAP = -2  ///< = -2.
81
    };
82

	
83
  private:
84
    std::vector<Pair> _data;
85
    Compare _comp;
86
    ItemIntMap &_iim;
87

	
88
  public:
89
    /// \brief Constructor.
90
    ///
91
    /// Constructor.
92
    /// \param map A map that assigns \c int values to the items.
93
    /// It is used internally to handle the cross references.
94
    /// The assigned value must be \c PRE_HEAP (<tt>-1</tt>) for each item.
95
    explicit FouraryHeap(ItemIntMap &map) : _iim(map) {}
96

	
97
    /// \brief Constructor.
98
    ///
99
    /// Constructor.
100
    /// \param map A map that assigns \c int values to the items.
101
    /// It is used internally to handle the cross references.
102
    /// The assigned value must be \c PRE_HEAP (<tt>-1</tt>) for each item.
103
    /// \param comp The function object used for comparing the priorities.
104
    FouraryHeap(ItemIntMap &map, const Compare &comp)
105
      : _iim(map), _comp(comp) {}
106

	
107
    /// \brief The number of items stored in the heap.
108
    ///
109
    /// This function returns the number of items stored in the heap.
110
    int size() const { return _data.size(); }
111

	
112
    /// \brief Check if the heap is empty.
113
    ///
114
    /// This function returns \c true if the heap is empty.
115
    bool empty() const { return _data.empty(); }
116

	
117
    /// \brief Make the heap empty.
118
    ///
119
    /// This functon makes the heap empty.
120
    /// It does not change the cross reference map. If you want to reuse
121
    /// a heap that is not surely empty, you should first clear it and
122
    /// then you should set the cross reference map to \c PRE_HEAP
123
    /// for each item.
124
    void clear() { _data.clear(); }
125

	
126
  private:
127
    static int parent(int i) { return (i-1)/4; }
128
    static int firstChild(int i) { return 4*i+1; }
129

	
130
    bool less(const Pair &p1, const Pair &p2) const {
131
      return _comp(p1.second, p2.second);
132
    }
133

	
134
    void bubbleUp(int hole, Pair p) {
135
      int par = parent(hole);
136
      while( hole>0 && less(p,_data[par]) ) {
137
        move(_data[par],hole);
138
        hole = par;
139
        par = parent(hole);
140
      }
141
      move(p, hole);
142
    }
143

	
144
    void bubbleDown(int hole, Pair p, int length) {
145
      if( length>1 ) {
146
        int child = firstChild(hole);
147
        while( child+3<length ) {
148
          int min=child;
149
          if( less(_data[++child], _data[min]) ) min=child;
150
          if( less(_data[++child], _data[min]) ) min=child;
151
          if( less(_data[++child], _data[min]) ) min=child;
152
          if( !less(_data[min], p) )
153
            goto ok;
154
          move(_data[min], hole);
155
          hole = min;
156
          child = firstChild(hole);
157
        }
158
        if ( child<length ) {
159
          int min = child;
160
          if( ++child<length && less(_data[child], _data[min]) ) min=child;
161
          if( ++child<length && less(_data[child], _data[min]) ) min=child;
162
          if( less(_data[min], p) ) {
163
            move(_data[min], hole);
164
            hole = min;
165
          }
166
        }
167
      }
168
    ok:
169
      move(p, hole);
170
    }
171

	
172
    void move(const Pair &p, int i) {
173
      _data[i] = p;
174
      _iim.set(p.first, i);
175
    }
176

	
177
  public:
178
    /// \brief Insert a pair of item and priority into the heap.
179
    ///
180
    /// This function inserts \c p.first to the heap with priority
181
    /// \c p.second.
182
    /// \param p The pair to insert.
183
    /// \pre \c p.first must not be stored in the heap.
184
    void push(const Pair &p) {
185
      int n = _data.size();
186
      _data.resize(n+1);
187
      bubbleUp(n, p);
188
    }
189

	
190
    /// \brief Insert an item into the heap with the given priority.
191
    ///
192
    /// This function inserts the given item into the heap with the
193
    /// given priority.
194
    /// \param i The item to insert.
195
    /// \param p The priority of the item.
196
    /// \pre \e i must not be stored in the heap.
197
    void push(const Item &i, const Prio &p) { push(Pair(i,p)); }
198

	
199
    /// \brief Return the item having minimum priority.
200
    ///
201
    /// This function returns the item having minimum priority.
202
    /// \pre The heap must be non-empty.
203
    Item top() const { return _data[0].first; }
204

	
205
    /// \brief The minimum priority.
206
    ///
207
    /// This function returns the minimum priority.
208
    /// \pre The heap must be non-empty.
209
    Prio prio() const { return _data[0].second; }
210

	
211
    /// \brief Remove the item having minimum priority.
212
    ///
213
    /// This function removes the item having minimum priority.
214
    /// \pre The heap must be non-empty.
215
    void pop() {
216
      int n = _data.size()-1;
217
      _iim.set(_data[0].first, POST_HEAP);
218
      if (n>0) bubbleDown(0, _data[n], n);
219
      _data.pop_back();
220
    }
221

	
222
    /// \brief Remove the given item from the heap.
223
    ///
224
    /// This function removes the given item from the heap if it is
225
    /// already stored.
226
    /// \param i The item to delete.
227
    /// \pre \e i must be in the heap.
228
    void erase(const Item &i) {
229
      int h = _iim[i];
230
      int n = _data.size()-1;
231
      _iim.set(_data[h].first, POST_HEAP);
232
      if( h<n ) {
233
        if( less(_data[parent(h)], _data[n]) )
234
          bubbleDown(h, _data[n], n);
235
        else
236
          bubbleUp(h, _data[n]);
237
      }
238
      _data.pop_back();
239
    }
240

	
241
    /// \brief The priority of the given item.
242
    ///
243
    /// This function returns the priority of the given item.
244
    /// \param i The item.
245
    /// \pre \e i must be in the heap.
246
    Prio operator[](const Item &i) const {
247
      int idx = _iim[i];
248
      return _data[idx].second;
249
    }
250

	
251
    /// \brief Set the priority of an item or insert it, if it is
252
    /// not stored in the heap.
253
    ///
254
    /// This method sets the priority of the given item if it is
255
    /// already stored in the heap. Otherwise it inserts the given
256
    /// item into the heap with the given priority.
257
    /// \param i The item.
258
    /// \param p The priority.
259
    void set(const Item &i, const Prio &p) {
260
      int idx = _iim[i];
261
      if( idx < 0 )
262
        push(i,p);
263
      else if( _comp(p, _data[idx].second) )
264
        bubbleUp(idx, Pair(i,p));
265
      else
266
        bubbleDown(idx, Pair(i,p), _data.size());
267
    }
268

	
269
    /// \brief Decrease the priority of an item to the given value.
270
    ///
271
    /// This function decreases the priority of an item to the given value.
272
    /// \param i The item.
273
    /// \param p The priority.
274
    /// \pre \e i must be stored in the heap with priority at least \e p.
275
    void decrease(const Item &i, const Prio &p) {
276
      int idx = _iim[i];
277
      bubbleUp(idx, Pair(i,p));
278
    }
279

	
280
    /// \brief Increase the priority of an item to the given value.
281
    ///
282
    /// This function increases the priority of an item to the given value.
283
    /// \param i The item.
284
    /// \param p The priority.
285
    /// \pre \e i must be stored in the heap with priority at most \e p.
286
    void increase(const Item &i, const Prio &p) {
287
      int idx = _iim[i];
288
      bubbleDown(idx, Pair(i,p), _data.size());
289
    }
290

	
291
    /// \brief Return the state of an item.
292
    ///
293
    /// This method returns \c PRE_HEAP if the given item has never
294
    /// been in the heap, \c IN_HEAP if it is in the heap at the moment,
295
    /// and \c POST_HEAP otherwise.
296
    /// In the latter case it is possible that the item will get back
297
    /// to the heap again.
298
    /// \param i The item.
299
    State state(const Item &i) const {
300
      int s = _iim[i];
301
      if (s>=0) s=0;
302
      return State(s);
303
    }
304

	
305
    /// \brief Set the state of an item in the heap.
306
    ///
307
    /// This function sets the state of the given item in the heap.
308
    /// It can be used to manually clear the heap when it is important
309
    /// to achive better time complexity.
310
    /// \param i The item.
311
    /// \param st The state. It should not be \c IN_HEAP.
312
    void state(const Item& i, State st) {
313
      switch (st) {
314
        case POST_HEAP:
315
        case PRE_HEAP:
316
          if (state(i) == IN_HEAP) erase(i);
317
          _iim[i] = st;
318
          break;
319
        case IN_HEAP:
320
          break;
321
      }
322
    }
323

	
324
    /// \brief Replace an item in the heap.
325
    ///
326
    /// This function replaces item \c i with item \c j.
327
    /// Item \c i must be in the heap, while \c j must be out of the heap.
328
    /// After calling this method, item \c i will be out of the
329
    /// heap and \c j will be in the heap with the same prioriority
330
    /// as item \c i had before.
331
    void replace(const Item& i, const Item& j) {
332
      int idx = _iim[i];
333
      _iim.set(i, _iim[j]);
334
      _iim.set(j, idx);
335
      _data[idx].first = j;
336
    }
337

	
338
  }; // class FouraryHeap
339

	
340
} // namespace lemon
341

	
342
#endif // LEMON_FOURARY_HEAP_H
Ignore white space 6 line context
1
/* -*- mode: C++; indent-tabs-mode: nil; -*-
2
 *
3
 * This file is a part of LEMON, a generic C++ optimization library.
4
 *
5
 * Copyright (C) 2003-2009
6
 * Egervary Jeno Kombinatorikus Optimalizalasi Kutatocsoport
7
 * (Egervary Research Group on Combinatorial Optimization, EGRES).
8
 *
9
 * Permission to use, modify and distribute this software is granted
10
 * provided that this copyright notice appears in all copies. For
11
 * precise terms see the accompanying LICENSE file.
12
 *
13
 * This software is provided "AS IS" with no warranty of any kind,
14
 * express or implied, and with no claim as to its suitability for any
15
 * purpose.
16
 *
17
 */
18

	
19
#ifndef LEMON_KARY_HEAP_H
20
#define LEMON_KARY_HEAP_H
21

	
22
///\ingroup heaps
23
///\file
24
///\brief Fourary heap implementation.
25

	
26
#include <vector>
27
#include <utility>
28
#include <functional>
29

	
30
namespace lemon {
31

	
32
  /// \ingroup heaps
33
  ///
34
  ///\brief K-ary heap data structure.
35
  ///
36
  /// This class implements the \e K-ary \e heap data structure.
37
  /// It fully conforms to the \ref concepts::Heap "heap concept".
38
  ///
39
  /// The \ref KaryHeap "K-ary heap" is a generalization of the
40
  /// \ref BinHeap "binary heap" structure, its nodes have at most
41
  /// \c K children, instead of two.
42
  /// \ref BinHeap and \ref FouraryHeap are specialized implementations
43
  /// of this structure for <tt>K=2</tt> and <tt>K=4</tt>, respectively.
44
  ///
45
  /// \tparam PR Type of the priorities of the items.
46
  /// \tparam IM A read-writable item map with \c int values, used
47
  /// internally to handle the cross references.
48
  /// \tparam K The degree of the heap, each node have at most \e K
49
  /// children. The default is 16. Powers of two are suggested to use
50
  /// so that the multiplications and divisions needed to traverse the
51
  /// nodes of the heap could be performed faster.
52
  /// \tparam CMP A functor class for comparing the priorities.
53
  /// The default is \c std::less<PR>.
54
  ///
55
  ///\sa BinHeap
56
  ///\sa FouraryHeap
57
#ifdef DOXYGEN
58
  template <typename PR, typename IM, int K, typename CMP>
59
#else
60
  template <typename PR, typename IM, int K = 16,
61
            typename CMP = std::less<PR> >
62
#endif
63
  class KaryHeap {
64
  public:
65
    /// Type of the item-int map.
66
    typedef IM ItemIntMap;
67
    /// Type of the priorities.
68
    typedef PR Prio;
69
    /// Type of the items stored in the heap.
70
    typedef typename ItemIntMap::Key Item;
71
    /// Type of the item-priority pairs.
72
    typedef std::pair<Item,Prio> Pair;
73
    /// Functor type for comparing the priorities.
74
    typedef CMP Compare;
75

	
76
    /// \brief Type to represent the states of the items.
77
    ///
78
    /// Each item has a state associated to it. It can be "in heap",
79
    /// "pre-heap" or "post-heap". The latter two are indifferent from the
80
    /// heap's point of view, but may be useful to the user.
81
    ///
82
    /// The item-int map must be initialized in such way that it assigns
83
    /// \c PRE_HEAP (<tt>-1</tt>) to any element to be put in the heap.
84
    enum State {
85
      IN_HEAP = 0,    ///< = 0.
86
      PRE_HEAP = -1,  ///< = -1.
87
      POST_HEAP = -2  ///< = -2.
88
    };
89

	
90
  private:
91
    std::vector<Pair> _data;
92
    Compare _comp;
93
    ItemIntMap &_iim;
94

	
95
  public:
96
    /// \brief Constructor.
97
    ///
98
    /// Constructor.
99
    /// \param map A map that assigns \c int values to the items.
100
    /// It is used internally to handle the cross references.
101
    /// The assigned value must be \c PRE_HEAP (<tt>-1</tt>) for each item.
102
    explicit KaryHeap(ItemIntMap &map) : _iim(map) {}
103

	
104
    /// \brief Constructor.
105
    ///
106
    /// Constructor.
107
    /// \param map A map that assigns \c int values to the items.
108
    /// It is used internally to handle the cross references.
109
    /// The assigned value must be \c PRE_HEAP (<tt>-1</tt>) for each item.
110
    /// \param comp The function object used for comparing the priorities.
111
    KaryHeap(ItemIntMap &map, const Compare &comp)
112
      : _iim(map), _comp(comp) {}
113

	
114
    /// \brief The number of items stored in the heap.
115
    ///
116
    /// This function returns the number of items stored in the heap.
117
    int size() const { return _data.size(); }
118

	
119
    /// \brief Check if the heap is empty.
120
    ///
121
    /// This function returns \c true if the heap is empty.
122
    bool empty() const { return _data.empty(); }
123

	
124
    /// \brief Make the heap empty.
125
    ///
126
    /// This functon makes the heap empty.
127
    /// It does not change the cross reference map. If you want to reuse
128
    /// a heap that is not surely empty, you should first clear it and
129
    /// then you should set the cross reference map to \c PRE_HEAP
130
    /// for each item.
131
    void clear() { _data.clear(); }
132

	
133
  private:
134
    int parent(int i) { return (i-1)/K; }
135
    int firstChild(int i) { return K*i+1; }
136

	
137
    bool less(const Pair &p1, const Pair &p2) const {
138
      return _comp(p1.second, p2.second);
139
    }
140

	
141
    void bubbleUp(int hole, Pair p) {
142
      int par = parent(hole);
143
      while( hole>0 && less(p,_data[par]) ) {
144
        move(_data[par],hole);
145
        hole = par;
146
        par = parent(hole);
147
      }
148
      move(p, hole);
149
    }
150

	
151
    void bubbleDown(int hole, Pair p, int length) {
152
      if( length>1 ) {
153
        int child = firstChild(hole);
154
        while( child+K<=length ) {
155
          int min=child;
156
          for (int i=1; i<K; ++i) {
157
            if( less(_data[child+i], _data[min]) )
158
              min=child+i;
159
          }
160
          if( !less(_data[min], p) )
161
            goto ok;
162
          move(_data[min], hole);
163
          hole = min;
164
          child = firstChild(hole);
165
        }
166
        if ( child<length ) {
167
          int min = child;
168
          while (++child < length) {
169
            if( less(_data[child], _data[min]) )
170
              min=child;
171
          }
172
          if( less(_data[min], p) ) {
173
            move(_data[min], hole);
174
            hole = min;
175
          }
176
        }
177
      }
178
    ok:
179
      move(p, hole);
180
    }
181

	
182
    void move(const Pair &p, int i) {
183
      _data[i] = p;
184
      _iim.set(p.first, i);
185
    }
186

	
187
  public:
188
    /// \brief Insert a pair of item and priority into the heap.
189
    ///
190
    /// This function inserts \c p.first to the heap with priority
191
    /// \c p.second.
192
    /// \param p The pair to insert.
193
    /// \pre \c p.first must not be stored in the heap.
194
    void push(const Pair &p) {
195
      int n = _data.size();
196
      _data.resize(n+1);
197
      bubbleUp(n, p);
198
    }
199

	
200
    /// \brief Insert an item into the heap with the given priority.
201
    ///
202
    /// This function inserts the given item into the heap with the
203
    /// given priority.
204
    /// \param i The item to insert.
205
    /// \param p The priority of the item.
206
    /// \pre \e i must not be stored in the heap.
207
    void push(const Item &i, const Prio &p) { push(Pair(i,p)); }
208

	
209
    /// \brief Return the item having minimum priority.
210
    ///
211
    /// This function returns the item having minimum priority.
212
    /// \pre The heap must be non-empty.
213
    Item top() const { return _data[0].first; }
214

	
215
    /// \brief The minimum priority.
216
    ///
217
    /// This function returns the minimum priority.
218
    /// \pre The heap must be non-empty.
219
    Prio prio() const { return _data[0].second; }
220

	
221
    /// \brief Remove the item having minimum priority.
222
    ///
223
    /// This function removes the item having minimum priority.
224
    /// \pre The heap must be non-empty.
225
    void pop() {
226
      int n = _data.size()-1;
227
      _iim.set(_data[0].first, POST_HEAP);
228
      if (n>0) bubbleDown(0, _data[n], n);
229
      _data.pop_back();
230
    }
231

	
232
    /// \brief Remove the given item from the heap.
233
    ///
234
    /// This function removes the given item from the heap if it is
235
    /// already stored.
236
    /// \param i The item to delete.
237
    /// \pre \e i must be in the heap.
238
    void erase(const Item &i) {
239
      int h = _iim[i];
240
      int n = _data.size()-1;
241
      _iim.set(_data[h].first, POST_HEAP);
242
      if( h<n ) {
243
        if( less(_data[parent(h)], _data[n]) )
244
          bubbleDown(h, _data[n], n);
245
        else
246
          bubbleUp(h, _data[n]);
247
      }
248
      _data.pop_back();
249
    }
250

	
251
    /// \brief The priority of the given item.
252
    ///
253
    /// This function returns the priority of the given item.
254
    /// \param i The item.
255
    /// \pre \e i must be in the heap.
256
    Prio operator[](const Item &i) const {
257
      int idx = _iim[i];
258
      return _data[idx].second;
259
    }
260

	
261
    /// \brief Set the priority of an item or insert it, if it is
262
    /// not stored in the heap.
263
    ///
264
    /// This method sets the priority of the given item if it is
265
    /// already stored in the heap. Otherwise it inserts the given
266
    /// item into the heap with the given priority.
267
    /// \param i The item.
268
    /// \param p The priority.
269
    void set(const Item &i, const Prio &p) {
270
      int idx = _iim[i];
271
      if( idx<0 )
272
        push(i,p);
273
      else if( _comp(p, _data[idx].second) )
274
        bubbleUp(idx, Pair(i,p));
275
      else
276
        bubbleDown(idx, Pair(i,p), _data.size());
277
    }
278

	
279
    /// \brief Decrease the priority of an item to the given value.
280
    ///
281
    /// This function decreases the priority of an item to the given value.
282
    /// \param i The item.
283
    /// \param p The priority.
284
    /// \pre \e i must be stored in the heap with priority at least \e p.
285
    void decrease(const Item &i, const Prio &p) {
286
      int idx = _iim[i];
287
      bubbleUp(idx, Pair(i,p));
288
    }
289

	
290
    /// \brief Increase the priority of an item to the given value.
291
    ///
292
    /// This function increases the priority of an item to the given value.
293
    /// \param i The item.
294
    /// \param p The priority.
295
    /// \pre \e i must be stored in the heap with priority at most \e p.
296
    void increase(const Item &i, const Prio &p) {
297
      int idx = _iim[i];
298
      bubbleDown(idx, Pair(i,p), _data.size());
299
    }
300

	
301
    /// \brief Return the state of an item.
302
    ///
303
    /// This method returns \c PRE_HEAP if the given item has never
304
    /// been in the heap, \c IN_HEAP if it is in the heap at the moment,
305
    /// and \c POST_HEAP otherwise.
306
    /// In the latter case it is possible that the item will get back
307
    /// to the heap again.
308
    /// \param i The item.
309
    State state(const Item &i) const {
310
      int s = _iim[i];
311
      if (s>=0) s=0;
312
      return State(s);
313
    }
314

	
315
    /// \brief Set the state of an item in the heap.
316
    ///
317
    /// This function sets the state of the given item in the heap.
318
    /// It can be used to manually clear the heap when it is important
319
    /// to achive better time complexity.
320
    /// \param i The item.
321
    /// \param st The state. It should not be \c IN_HEAP.
322
    void state(const Item& i, State st) {
323
      switch (st) {
324
        case POST_HEAP:
325
        case PRE_HEAP:
326
          if (state(i) == IN_HEAP) erase(i);
327
          _iim[i] = st;
328
          break;
329
        case IN_HEAP:
330
          break;
331
      }
332
    }
333

	
334
    /// \brief Replace an item in the heap.
335
    ///
336
    /// This function replaces item \c i with item \c j.
337
    /// Item \c i must be in the heap, while \c j must be out of the heap.
338
    /// After calling this method, item \c i will be out of the
339
    /// heap and \c j will be in the heap with the same prioriority
340
    /// as item \c i had before.
341
    void replace(const Item& i, const Item& j) {
342
      int idx=_iim[i];
343
      _iim.set(i, _iim[j]);
344
      _iim.set(j, idx);
345
      _data[idx].first=j;
346
    }
347

	
348
  }; // class KaryHeap
349

	
350
} // namespace lemon
351

	
352
#endif // LEMON_KARY_HEAP_H
Ignore white space 6 line context
... ...
@@ -133,540 +133,583 @@
133 133
int algorithm1(const ListDigraph& g) {
134 134
  ReverseDigraph<const ListDigraph> rg(g);
135 135
  return algorithm2(rg);
136 136
}
137 137
\endcode
138 138
*/
139 139

	
140 140
/**
141 141
@defgroup maps Maps
142 142
@ingroup datas
143 143
\brief Map structures implemented in LEMON.
144 144

	
145 145
This group contains the map structures implemented in LEMON.
146 146

	
147 147
LEMON provides several special purpose maps and map adaptors that e.g. combine
148 148
new maps from existing ones.
149 149

	
150 150
<b>See also:</b> \ref map_concepts "Map Concepts".
151 151
*/
152 152

	
153 153
/**
154 154
@defgroup graph_maps Graph Maps
155 155
@ingroup maps
156 156
\brief Special graph-related maps.
157 157

	
158 158
This group contains maps that are specifically designed to assign
159 159
values to the nodes and arcs/edges of graphs.
160 160

	
161 161
If you are looking for the standard graph maps (\c NodeMap, \c ArcMap,
162 162
\c EdgeMap), see the \ref graph_concepts "Graph Structure Concepts".
163 163
*/
164 164

	
165 165
/**
166 166
\defgroup map_adaptors Map Adaptors
167 167
\ingroup maps
168 168
\brief Tools to create new maps from existing ones
169 169

	
170 170
This group contains map adaptors that are used to create "implicit"
171 171
maps from other maps.
172 172

	
173 173
Most of them are \ref concepts::ReadMap "read-only maps".
174 174
They can make arithmetic and logical operations between one or two maps
175 175
(negation, shifting, addition, multiplication, logical 'and', 'or',
176 176
'not' etc.) or e.g. convert a map to another one of different Value type.
177 177

	
178 178
The typical usage of this classes is passing implicit maps to
179 179
algorithms.  If a function type algorithm is called then the function
180 180
type map adaptors can be used comfortable. For example let's see the
181 181
usage of map adaptors with the \c graphToEps() function.
182 182
\code
183 183
  Color nodeColor(int deg) {
184 184
    if (deg >= 2) {
185 185
      return Color(0.5, 0.0, 0.5);
186 186
    } else if (deg == 1) {
187 187
      return Color(1.0, 0.5, 1.0);
188 188
    } else {
189 189
      return Color(0.0, 0.0, 0.0);
190 190
    }
191 191
  }
192 192

	
193 193
  Digraph::NodeMap<int> degree_map(graph);
194 194

	
195 195
  graphToEps(graph, "graph.eps")
196 196
    .coords(coords).scaleToA4().undirected()
197 197
    .nodeColors(composeMap(functorToMap(nodeColor), degree_map))
198 198
    .run();
199 199
\endcode
200 200
The \c functorToMap() function makes an \c int to \c Color map from the
201 201
\c nodeColor() function. The \c composeMap() compose the \c degree_map
202 202
and the previously created map. The composed map is a proper function to
203 203
get the color of each node.
204 204

	
205 205
The usage with class type algorithms is little bit harder. In this
206 206
case the function type map adaptors can not be used, because the
207 207
function map adaptors give back temporary objects.
208 208
\code
209 209
  Digraph graph;
210 210

	
211 211
  typedef Digraph::ArcMap<double> DoubleArcMap;
212 212
  DoubleArcMap length(graph);
213 213
  DoubleArcMap speed(graph);
214 214

	
215 215
  typedef DivMap<DoubleArcMap, DoubleArcMap> TimeMap;
216 216
  TimeMap time(length, speed);
217 217

	
218 218
  Dijkstra<Digraph, TimeMap> dijkstra(graph, time);
219 219
  dijkstra.run(source, target);
220 220
\endcode
221 221
We have a length map and a maximum speed map on the arcs of a digraph.
222 222
The minimum time to pass the arc can be calculated as the division of
223 223
the two maps which can be done implicitly with the \c DivMap template
224 224
class. We use the implicit minimum time map as the length map of the
225 225
\c Dijkstra algorithm.
226 226
*/
227 227

	
228 228
/**
229
@defgroup matrices Matrices
230
@ingroup datas
231
\brief Two dimensional data storages implemented in LEMON.
232

	
233
This group contains two dimensional data storages implemented in LEMON.
234
*/
235

	
236
/**
237 229
@defgroup paths Path Structures
238 230
@ingroup datas
239 231
\brief %Path structures implemented in LEMON.
240 232

	
241 233
This group contains the path structures implemented in LEMON.
242 234

	
243 235
LEMON provides flexible data structures to work with paths.
244 236
All of them have similar interfaces and they can be copied easily with
245 237
assignment operators and copy constructors. This makes it easy and
246 238
efficient to have e.g. the Dijkstra algorithm to store its result in
247 239
any kind of path structure.
248 240

	
249
\sa lemon::concepts::Path
241
\sa \ref concepts::Path "Path concept"
242
*/
243

	
244
/**
245
@defgroup heaps Heap Structures
246
@ingroup datas
247
\brief %Heap structures implemented in LEMON.
248

	
249
This group contains the heap structures implemented in LEMON.
250

	
251
LEMON provides several heap classes. They are efficient implementations
252
of the abstract data type \e priority \e queue. They store items with
253
specified values called \e priorities in such a way that finding and
254
removing the item with minimum priority are efficient.
255
The basic operations are adding and erasing items, changing the priority
256
of an item, etc.
257

	
258
Heaps are crucial in several algorithms, such as Dijkstra and Prim.
259
The heap implementations have the same interface, thus any of them can be
260
used easily in such algorithms.
261

	
262
\sa \ref concepts::Heap "Heap concept"
263
*/
264

	
265
/**
266
@defgroup matrices Matrices
267
@ingroup datas
268
\brief Two dimensional data storages implemented in LEMON.
269

	
270
This group contains two dimensional data storages implemented in LEMON.
250 271
*/
251 272

	
252 273
/**
253 274
@defgroup auxdat Auxiliary Data Structures
254 275
@ingroup datas
255 276
\brief Auxiliary data structures implemented in LEMON.
256 277

	
257 278
This group contains some data structures implemented in LEMON in
258 279
order to make it easier to implement combinatorial algorithms.
259 280
*/
260 281

	
261 282
/**
283
@defgroup geomdat Geometric Data Structures
284
@ingroup auxdat
285
\brief Geometric data structures implemented in LEMON.
286

	
287
This group contains geometric data structures implemented in LEMON.
288

	
289
 - \ref lemon::dim2::Point "dim2::Point" implements a two dimensional
290
   vector with the usual operations.
291
 - \ref lemon::dim2::Box "dim2::Box" can be used to determine the
292
   rectangular bounding box of a set of \ref lemon::dim2::Point
293
   "dim2::Point"'s.
294
*/
295

	
296
/**
297
@defgroup matrices Matrices
298
@ingroup auxdat
299
\brief Two dimensional data storages implemented in LEMON.
300

	
301
This group contains two dimensional data storages implemented in LEMON.
302
*/
303

	
304
/**
262 305
@defgroup algs Algorithms
263 306
\brief This group contains the several algorithms
264 307
implemented in LEMON.
265 308

	
266 309
This group contains the several algorithms
267 310
implemented in LEMON.
268 311
*/
269 312

	
270 313
/**
271 314
@defgroup search Graph Search
272 315
@ingroup algs
273 316
\brief Common graph search algorithms.
274 317

	
275 318
This group contains the common graph search algorithms, namely
276 319
\e breadth-first \e search (BFS) and \e depth-first \e search (DFS).
277 320
*/
278 321

	
279 322
/**
280 323
@defgroup shortest_path Shortest Path Algorithms
281 324
@ingroup algs
282 325
\brief Algorithms for finding shortest paths.
283 326

	
284 327
This group contains the algorithms for finding shortest paths in digraphs.
285 328

	
286 329
 - \ref Dijkstra algorithm for finding shortest paths from a source node
287 330
   when all arc lengths are non-negative.
288 331
 - \ref BellmanFord "Bellman-Ford" algorithm for finding shortest paths
289 332
   from a source node when arc lenghts can be either positive or negative,
290 333
   but the digraph should not contain directed cycles with negative total
291 334
   length.
292 335
 - \ref FloydWarshall "Floyd-Warshall" and \ref Johnson "Johnson" algorithms
293 336
   for solving the \e all-pairs \e shortest \e paths \e problem when arc
294 337
   lenghts can be either positive or negative, but the digraph should
295 338
   not contain directed cycles with negative total length.
296 339
 - \ref Suurballe A successive shortest path algorithm for finding
297 340
   arc-disjoint paths between two nodes having minimum total length.
298 341
*/
299 342

	
300 343
/**
344
@defgroup spantree Minimum Spanning Tree Algorithms
345
@ingroup algs
346
\brief Algorithms for finding minimum cost spanning trees and arborescences.
347

	
348
This group contains the algorithms for finding minimum cost spanning
349
trees and arborescences.
350
*/
351

	
352
/**
301 353
@defgroup max_flow Maximum Flow Algorithms
302 354
@ingroup algs
303 355
\brief Algorithms for finding maximum flows.
304 356

	
305 357
This group contains the algorithms for finding maximum flows and
306 358
feasible circulations.
307 359

	
308 360
The \e maximum \e flow \e problem is to find a flow of maximum value between
309 361
a single source and a single target. Formally, there is a \f$G=(V,A)\f$
310 362
digraph, a \f$cap: A\rightarrow\mathbf{R}^+_0\f$ capacity function and
311 363
\f$s, t \in V\f$ source and target nodes.
312 364
A maximum flow is an \f$f: A\rightarrow\mathbf{R}^+_0\f$ solution of the
313 365
following optimization problem.
314 366

	
315 367
\f[ \max\sum_{sv\in A} f(sv) - \sum_{vs\in A} f(vs) \f]
316 368
\f[ \sum_{uv\in A} f(uv) = \sum_{vu\in A} f(vu)
317 369
    \quad \forall u\in V\setminus\{s,t\} \f]
318 370
\f[ 0 \leq f(uv) \leq cap(uv) \quad \forall uv\in A \f]
319 371

	
320 372
LEMON contains several algorithms for solving maximum flow problems:
321 373
- \ref EdmondsKarp Edmonds-Karp algorithm.
322 374
- \ref Preflow Goldberg-Tarjan's preflow push-relabel algorithm.
323 375
- \ref DinitzSleatorTarjan Dinitz's blocking flow algorithm with dynamic trees.
324 376
- \ref GoldbergTarjan Preflow push-relabel algorithm with dynamic trees.
325 377

	
326 378
In most cases the \ref Preflow "Preflow" algorithm provides the
327 379
fastest method for computing a maximum flow. All implementations
328 380
also provide functions to query the minimum cut, which is the dual
329 381
problem of maximum flow.
330 382

	
331 383
\ref Circulation is a preflow push-relabel algorithm implemented directly 
332 384
for finding feasible circulations, which is a somewhat different problem,
333 385
but it is strongly related to maximum flow.
334 386
For more information, see \ref Circulation.
335 387
*/
336 388

	
337 389
/**
338 390
@defgroup min_cost_flow_algs Minimum Cost Flow Algorithms
339 391
@ingroup algs
340 392

	
341 393
\brief Algorithms for finding minimum cost flows and circulations.
342 394

	
343 395
This group contains the algorithms for finding minimum cost flows and
344 396
circulations. For more information about this problem and its dual
345 397
solution see \ref min_cost_flow "Minimum Cost Flow Problem".
346 398

	
347 399
LEMON contains several algorithms for this problem.
348 400
 - \ref NetworkSimplex Primal Network Simplex algorithm with various
349 401
   pivot strategies.
350 402
 - \ref CostScaling Push-Relabel and Augment-Relabel algorithms based on
351 403
   cost scaling.
352 404
 - \ref CapacityScaling Successive Shortest %Path algorithm with optional
353 405
   capacity scaling.
354 406
 - \ref CancelAndTighten The Cancel and Tighten algorithm.
355 407
 - \ref CycleCanceling Cycle-Canceling algorithms.
356 408

	
357 409
In general NetworkSimplex is the most efficient implementation,
358 410
but in special cases other algorithms could be faster.
359 411
For example, if the total supply and/or capacities are rather small,
360 412
CapacityScaling is usually the fastest algorithm (without effective scaling).
361 413
*/
362 414

	
363 415
/**
364 416
@defgroup min_cut Minimum Cut Algorithms
365 417
@ingroup algs
366 418

	
367 419
\brief Algorithms for finding minimum cut in graphs.
368 420

	
369 421
This group contains the algorithms for finding minimum cut in graphs.
370 422

	
371 423
The \e minimum \e cut \e problem is to find a non-empty and non-complete
372 424
\f$X\f$ subset of the nodes with minimum overall capacity on
373 425
outgoing arcs. Formally, there is a \f$G=(V,A)\f$ digraph, a
374 426
\f$cap: A\rightarrow\mathbf{R}^+_0\f$ capacity function. The minimum
375 427
cut is the \f$X\f$ solution of the next optimization problem:
376 428

	
377 429
\f[ \min_{X \subset V, X\not\in \{\emptyset, V\}}
378
    \sum_{uv\in A, u\in X, v\not\in X}cap(uv) \f]
430
    \sum_{uv\in A: u\in X, v\not\in X}cap(uv) \f]
379 431

	
380 432
LEMON contains several algorithms related to minimum cut problems:
381 433

	
382 434
- \ref HaoOrlin "Hao-Orlin algorithm" for calculating minimum cut
383 435
  in directed graphs.
384 436
- \ref NagamochiIbaraki "Nagamochi-Ibaraki algorithm" for
385 437
  calculating minimum cut in undirected graphs.
386 438
- \ref GomoryHu "Gomory-Hu tree computation" for calculating
387 439
  all-pairs minimum cut in undirected graphs.
388 440

	
389 441
If you want to find minimum cut just between two distinict nodes,
390 442
see the \ref max_flow "maximum flow problem".
391 443
*/
392 444

	
393 445
/**
394
@defgroup graph_properties Connectivity and Other Graph Properties
395
@ingroup algs
396
\brief Algorithms for discovering the graph properties
397

	
398
This group contains the algorithms for discovering the graph properties
399
like connectivity, bipartiteness, euler property, simplicity etc.
400

	
401
\image html edge_biconnected_components.png
402
\image latex edge_biconnected_components.eps "bi-edge-connected components" width=\textwidth
403
*/
404

	
405
/**
406
@defgroup planar Planarity Embedding and Drawing
407
@ingroup algs
408
\brief Algorithms for planarity checking, embedding and drawing
409

	
410
This group contains the algorithms for planarity checking,
411
embedding and drawing.
412

	
413
\image html planar.png
414
\image latex planar.eps "Plane graph" width=\textwidth
415
*/
416

	
417
/**
418 446
@defgroup matching Matching Algorithms
419 447
@ingroup algs
420 448
\brief Algorithms for finding matchings in graphs and bipartite graphs.
421 449

	
422 450
This group contains the algorithms for calculating
423 451
matchings in graphs and bipartite graphs. The general matching problem is
424 452
finding a subset of the edges for which each node has at most one incident
425 453
edge.
426 454

	
427 455
There are several different algorithms for calculate matchings in
428 456
graphs.  The matching problems in bipartite graphs are generally
429 457
easier than in general graphs. The goal of the matching optimization
430 458
can be finding maximum cardinality, maximum weight or minimum cost
431 459
matching. The search can be constrained to find perfect or
432 460
maximum cardinality matching.
433 461

	
434 462
The matching algorithms implemented in LEMON:
435 463
- \ref MaxBipartiteMatching Hopcroft-Karp augmenting path algorithm
436 464
  for calculating maximum cardinality matching in bipartite graphs.
437 465
- \ref PrBipartiteMatching Push-relabel algorithm
438 466
  for calculating maximum cardinality matching in bipartite graphs.
439 467
- \ref MaxWeightedBipartiteMatching
440 468
  Successive shortest path algorithm for calculating maximum weighted
441 469
  matching and maximum weighted bipartite matching in bipartite graphs.
442 470
- \ref MinCostMaxBipartiteMatching
443 471
  Successive shortest path algorithm for calculating minimum cost maximum
444 472
  matching in bipartite graphs.
445 473
- \ref MaxMatching Edmond's blossom shrinking algorithm for calculating
446 474
  maximum cardinality matching in general graphs.
447 475
- \ref MaxWeightedMatching Edmond's blossom shrinking algorithm for calculating
448 476
  maximum weighted matching in general graphs.
449 477
- \ref MaxWeightedPerfectMatching
450 478
  Edmond's blossom shrinking algorithm for calculating maximum weighted
451 479
  perfect matching in general graphs.
452 480

	
453 481
\image html bipartite_matching.png
454 482
\image latex bipartite_matching.eps "Bipartite Matching" width=\textwidth
455 483
*/
456 484

	
457 485
/**
458
@defgroup spantree Minimum Spanning Tree Algorithms
486
@defgroup graph_properties Connectivity and Other Graph Properties
459 487
@ingroup algs
460
\brief Algorithms for finding minimum cost spanning trees and arborescences.
488
\brief Algorithms for discovering the graph properties
461 489

	
462
This group contains the algorithms for finding minimum cost spanning
463
trees and arborescences.
490
This group contains the algorithms for discovering the graph properties
491
like connectivity, bipartiteness, euler property, simplicity etc.
492

	
493
\image html connected_components.png
494
\image latex connected_components.eps "Connected components" width=\textwidth
495
*/
496

	
497
/**
498
@defgroup planar Planarity Embedding and Drawing
499
@ingroup algs
500
\brief Algorithms for planarity checking, embedding and drawing
501

	
502
This group contains the algorithms for planarity checking,
503
embedding and drawing.
504

	
505
\image html planar.png
506
\image latex planar.eps "Plane graph" width=\textwidth
507
*/
508

	
509
/**
510
@defgroup approx Approximation Algorithms
511
@ingroup algs
512
\brief Approximation algorithms.
513

	
514
This group contains the approximation and heuristic algorithms
515
implemented in LEMON.
464 516
*/
465 517

	
466 518
/**
467 519
@defgroup auxalg Auxiliary Algorithms
468 520
@ingroup algs
469 521
\brief Auxiliary algorithms implemented in LEMON.
470 522

	
471 523
This group contains some algorithms implemented in LEMON
472 524
in order to make it easier to implement complex algorithms.
473 525
*/
474 526

	
475 527
/**
476
@defgroup approx Approximation Algorithms
477
@ingroup algs
478
\brief Approximation algorithms.
479

	
480
This group contains the approximation and heuristic algorithms
481
implemented in LEMON.
482
*/
483

	
484
/**
485 528
@defgroup gen_opt_group General Optimization Tools
486 529
\brief This group contains some general optimization frameworks
487 530
implemented in LEMON.
488 531

	
489 532
This group contains some general optimization frameworks
490 533
implemented in LEMON.
491 534
*/
492 535

	
493 536
/**
494 537
@defgroup lp_group Lp and Mip Solvers
495 538
@ingroup gen_opt_group
496 539
\brief Lp and Mip solver interfaces for LEMON.
497 540

	
498 541
This group contains Lp and Mip solver interfaces for LEMON. The
499 542
various LP solvers could be used in the same manner with this
500 543
interface.
501 544
*/
502 545

	
503 546
/**
504 547
@defgroup lp_utils Tools for Lp and Mip Solvers
505 548
@ingroup lp_group
506 549
\brief Helper tools to the Lp and Mip solvers.
507 550

	
508 551
This group adds some helper tools to general optimization framework
509 552
implemented in LEMON.
510 553
*/
511 554

	
512 555
/**
513 556
@defgroup metah Metaheuristics
514 557
@ingroup gen_opt_group
515 558
\brief Metaheuristics for LEMON library.
516 559

	
517 560
This group contains some metaheuristic optimization tools.
518 561
*/
519 562

	
520 563
/**
521 564
@defgroup utils Tools and Utilities
522 565
\brief Tools and utilities for programming in LEMON
523 566

	
524 567
Tools and utilities for programming in LEMON.
525 568
*/
526 569

	
527 570
/**
528 571
@defgroup gutils Basic Graph Utilities
529 572
@ingroup utils
530 573
\brief Simple basic graph utilities.
531 574

	
532 575
This group contains some simple basic graph utilities.
533 576
*/
534 577

	
535 578
/**
536 579
@defgroup misc Miscellaneous Tools
537 580
@ingroup utils
538 581
\brief Tools for development, debugging and testing.
539 582

	
540 583
This group contains several useful tools for development,
541 584
debugging and testing.
542 585
*/
543 586

	
544 587
/**
545 588
@defgroup timecount Time Measuring and Counting
546 589
@ingroup misc
547 590
\brief Simple tools for measuring the performance of algorithms.
548 591

	
549 592
This group contains simple tools for measuring the performance
550 593
of algorithms.
551 594
*/
552 595

	
553 596
/**
554 597
@defgroup exceptions Exceptions
555 598
@ingroup utils
556 599
\brief Exceptions defined in LEMON.
557 600

	
558 601
This group contains the exceptions defined in LEMON.
559 602
*/
560 603

	
561 604
/**
562 605
@defgroup io_group Input-Output
563 606
\brief Graph Input-Output methods
564 607

	
565 608
This group contains the tools for importing and exporting graphs
566 609
and graph related data. Now it supports the \ref lgf-format
567 610
"LEMON Graph Format", the \c DIMACS format and the encapsulated
568 611
postscript (EPS) format.
569 612
*/
570 613

	
571 614
/**
572 615
@defgroup lemon_io LEMON Graph Format
573 616
@ingroup io_group
574 617
\brief Reading and writing LEMON Graph Format.
575 618

	
576 619
This group contains methods for reading and writing
577 620
\ref lgf-format "LEMON Graph Format".
578 621
*/
579 622

	
580 623
/**
581 624
@defgroup eps_io Postscript Exporting
582 625
@ingroup io_group
583 626
\brief General \c EPS drawer and graph exporter
584 627

	
585 628
This group contains general \c EPS drawing methods and special
586 629
graph exporting tools.
587 630
*/
588 631

	
589 632
/**
590
@defgroup dimacs_group DIMACS format
633
@defgroup dimacs_group DIMACS Format
591 634
@ingroup io_group
592 635
\brief Read and write files in DIMACS format
593 636

	
594 637
Tools to read a digraph from or write it to a file in DIMACS format data.
595 638
*/
596 639

	
597 640
/**
598 641
@defgroup nauty_group NAUTY Format
599 642
@ingroup io_group
600 643
\brief Read \e Nauty format
601 644

	
602 645
Tool to read graphs from \e Nauty format data.
603 646
*/
604 647

	
605 648
/**
606 649
@defgroup concept Concepts
607 650
\brief Skeleton classes and concept checking classes
608 651

	
609 652
This group contains the data/algorithm skeletons and concept checking
610 653
classes implemented in LEMON.
611 654

	
612 655
The purpose of the classes in this group is fourfold.
613 656

	
614 657
- These classes contain the documentations of the %concepts. In order
615 658
  to avoid document multiplications, an implementation of a concept
616 659
  simply refers to the corresponding concept class.
617 660

	
618 661
- These classes declare every functions, <tt>typedef</tt>s etc. an
619 662
  implementation of the %concepts should provide, however completely
620 663
  without implementations and real data structures behind the
621 664
  interface. On the other hand they should provide nothing else. All
622 665
  the algorithms working on a data structure meeting a certain concept
623 666
  should compile with these classes. (Though it will not run properly,
624 667
  of course.) In this way it is easily to check if an algorithm
625 668
  doesn't use any extra feature of a certain implementation.
626 669

	
627 670
- The concept descriptor classes also provide a <em>checker class</em>
628 671
  that makes it possible to check whether a certain implementation of a
629 672
  concept indeed provides all the required features.
630 673

	
631 674
- Finally, They can serve as a skeleton of a new implementation of a concept.
632 675
*/
633 676

	
634 677
/**
635 678
@defgroup graph_concepts Graph Structure Concepts
636 679
@ingroup concept
637 680
\brief Skeleton and concept checking classes for graph structures
638 681

	
639 682
This group contains the skeletons and concept checking classes of LEMON's
640 683
graph structures and helper classes used to implement these.
641 684
*/
642 685

	
643 686
/**
644 687
@defgroup map_concepts Map Concepts
645 688
@ingroup concept
646 689
\brief Skeleton and concept checking classes for maps
647 690

	
648 691
This group contains the skeletons and concept checking classes of maps.
649 692
*/
650 693

	
651 694
/**
695
@defgroup tools Standalone Utility Applications
696

	
697
Some utility applications are listed here.
698

	
699
The standard compilation procedure (<tt>./configure;make</tt>) will compile
700
them, as well.
701
*/
702

	
703
/**
652 704
\anchor demoprograms
653 705

	
654 706
@defgroup demos Demo Programs
655 707

	
656 708
Some demo programs are listed here. Their full source codes can be found in
657 709
the \c demo subdirectory of the source tree.
658 710

	
659 711
In order to compile them, use the <tt>make demo</tt> or the
660 712
<tt>make check</tt> commands.
661 713
*/
662 714

	
663
/**
664
@defgroup tools Standalone Utility Applications
665

	
666
Some utility applications are listed here.
667

	
668
The standard compilation procedure (<tt>./configure;make</tt>) will compile
669
them, as well.
670
*/
671

	
672 715
}
Ignore white space 6 line context
1 1
EXTRA_DIST += \
2 2
	lemon/lemon.pc.in \
3 3
	lemon/CMakeLists.txt \
4 4
	lemon/config.h.cmake
5 5

	
6 6
pkgconfig_DATA += lemon/lemon.pc
7 7

	
8 8
lib_LTLIBRARIES += lemon/libemon.la
9 9

	
10 10
lemon_libemon_la_SOURCES = \
11 11
	lemon/arg_parser.cc \
12 12
	lemon/base.cc \
13 13
	lemon/color.cc \
14 14
	lemon/lp_base.cc \
15 15
	lemon/lp_skeleton.cc \
16 16
	lemon/random.cc \
17 17
	lemon/bits/windows.cc
18 18

	
19 19
nodist_lemon_HEADERS = lemon/config.h	
20 20

	
21 21
lemon_libemon_la_CXXFLAGS = \
22 22
	$(AM_CXXFLAGS) \
23 23
	$(GLPK_CFLAGS) \
24 24
	$(CPLEX_CFLAGS) \
25 25
	$(SOPLEX_CXXFLAGS) \
26 26
	$(CLP_CXXFLAGS) \
27 27
	$(CBC_CXXFLAGS)
28 28

	
29 29
lemon_libemon_la_LDFLAGS = \
30 30
	$(GLPK_LIBS) \
31 31
	$(CPLEX_LIBS) \
32 32
	$(SOPLEX_LIBS) \
33 33
	$(CLP_LIBS) \
34 34
	$(CBC_LIBS)
35 35

	
36 36
if HAVE_GLPK
37 37
lemon_libemon_la_SOURCES += lemon/glpk.cc
38 38
endif
39 39

	
40 40
if HAVE_CPLEX
41 41
lemon_libemon_la_SOURCES += lemon/cplex.cc
42 42
endif
43 43

	
44 44
if HAVE_SOPLEX
45 45
lemon_libemon_la_SOURCES += lemon/soplex.cc
46 46
endif
47 47

	
48 48
if HAVE_CLP
49 49
lemon_libemon_la_SOURCES += lemon/clp.cc
50 50
endif
51 51

	
52 52
if HAVE_CBC
53 53
lemon_libemon_la_SOURCES += lemon/cbc.cc
54 54
endif
55 55

	
56 56
lemon_HEADERS += \
57 57
	lemon/adaptors.h \
58 58
	lemon/arg_parser.h \
59 59
	lemon/assert.h \
60
	lemon/bellman_ford.h \
60 61
	lemon/bfs.h \
61 62
	lemon/bin_heap.h \
63
	lemon/binom_heap.h \
64
	lemon/bucket_heap.h \
62 65
	lemon/cbc.h \
63 66
	lemon/circulation.h \
64 67
	lemon/clp.h \
65 68
	lemon/color.h \
66 69
	lemon/concept_check.h \
67 70
	lemon/connectivity.h \
68 71
	lemon/counter.h \
69 72
	lemon/core.h \
70 73
	lemon/cplex.h \
71 74
	lemon/dfs.h \
72 75
	lemon/dijkstra.h \
73 76
	lemon/dim2.h \
74 77
	lemon/dimacs.h \
75 78
	lemon/edge_set.h \
76 79
	lemon/elevator.h \
77 80
	lemon/error.h \
78 81
	lemon/euler.h \
82
	lemon/fib_heap.h \
83
	lemon/fourary_heap.h \
79 84
	lemon/full_graph.h \
80 85
	lemon/glpk.h \
81 86
	lemon/gomory_hu.h \
82 87
	lemon/graph_to_eps.h \
83 88
	lemon/grid_graph.h \
84 89
	lemon/hypercube_graph.h \
90
	lemon/kary_heap.h \
85 91
	lemon/kruskal.h \
86 92
	lemon/hao_orlin.h \
87 93
	lemon/lgf_reader.h \
88 94
	lemon/lgf_writer.h \
89 95
	lemon/list_graph.h \
90 96
	lemon/lp.h \
91 97
	lemon/lp_base.h \
92 98
	lemon/lp_skeleton.h \
93
	lemon/list_graph.h \
94 99
	lemon/maps.h \
95 100
	lemon/matching.h \
96 101
	lemon/math.h \
97 102
	lemon/min_cost_arborescence.h \
98 103
	lemon/nauty_reader.h \
99 104
	lemon/network_simplex.h \
105
	lemon/pairing_heap.h \
100 106
	lemon/path.h \
101 107
	lemon/preflow.h \
108
	lemon/radix_heap.h \
102 109
	lemon/radix_sort.h \
103 110
	lemon/random.h \
104 111
	lemon/smart_graph.h \
105 112
	lemon/soplex.h \
106 113
	lemon/suurballe.h \
107 114
	lemon/time_measure.h \
108 115
	lemon/tolerance.h \
109 116
	lemon/unionfind.h \
110 117
	lemon/bits/windows.h
111 118

	
112 119
bits_HEADERS += \
113 120
	lemon/bits/alteration_notifier.h \
114 121
	lemon/bits/array_map.h \
115 122
	lemon/bits/bezier.h \
116 123
	lemon/bits/default_map.h \
117 124
	lemon/bits/edge_set_extender.h \
118 125
	lemon/bits/enable_if.h \
119 126
	lemon/bits/graph_adaptor_extender.h \
120 127
	lemon/bits/graph_extender.h \
121 128
	lemon/bits/map_extender.h \
122 129
	lemon/bits/path_dump.h \
123 130
	lemon/bits/solver_bits.h \
124 131
	lemon/bits/traits.h \
125 132
	lemon/bits/variant.h \
126 133
	lemon/bits/vector_map.h
127 134

	
128 135
concept_HEADERS += \
129 136
	lemon/concepts/digraph.h \
130 137
	lemon/concepts/graph.h \
131 138
	lemon/concepts/graph_components.h \
132 139
	lemon/concepts/heap.h \
133 140
	lemon/concepts/maps.h \
134 141
	lemon/concepts/path.h
Ignore white space 6 line context
... ...
@@ -321,194 +321,194 @@
321 321
    ///\param g The digraph the algorithm runs on.
322 322
    Bfs(const Digraph &g) :
323 323
      G(&g),
324 324
      _pred(NULL), local_pred(false),
325 325
      _dist(NULL), local_dist(false),
326 326
      _reached(NULL), local_reached(false),
327 327
      _processed(NULL), local_processed(false)
328 328
    { }
329 329

	
330 330
    ///Destructor.
331 331
    ~Bfs()
332 332
    {
333 333
      if(local_pred) delete _pred;
334 334
      if(local_dist) delete _dist;
335 335
      if(local_reached) delete _reached;
336 336
      if(local_processed) delete _processed;
337 337
    }
338 338

	
339 339
    ///Sets the map that stores the predecessor arcs.
340 340

	
341 341
    ///Sets the map that stores the predecessor arcs.
342 342
    ///If you don't use this function before calling \ref run(Node) "run()"
343 343
    ///or \ref init(), an instance will be allocated automatically.
344 344
    ///The destructor deallocates this automatically allocated map,
345 345
    ///of course.
346 346
    ///\return <tt> (*this) </tt>
347 347
    Bfs &predMap(PredMap &m)
348 348
    {
349 349
      if(local_pred) {
350 350
        delete _pred;
351 351
        local_pred=false;
352 352
      }
353 353
      _pred = &m;
354 354
      return *this;
355 355
    }
356 356

	
357 357
    ///Sets the map that indicates which nodes are reached.
358 358

	
359 359
    ///Sets the map that indicates which nodes are reached.
360 360
    ///If you don't use this function before calling \ref run(Node) "run()"
361 361
    ///or \ref init(), an instance will be allocated automatically.
362 362
    ///The destructor deallocates this automatically allocated map,
363 363
    ///of course.
364 364
    ///\return <tt> (*this) </tt>
365 365
    Bfs &reachedMap(ReachedMap &m)
366 366
    {
367 367
      if(local_reached) {
368 368
        delete _reached;
369 369
        local_reached=false;
370 370
      }
371 371
      _reached = &m;
372 372
      return *this;
373 373
    }
374 374

	
375 375
    ///Sets the map that indicates which nodes are processed.
376 376

	
377 377
    ///Sets the map that indicates which nodes are processed.
378 378
    ///If you don't use this function before calling \ref run(Node) "run()"
379 379
    ///or \ref init(), an instance will be allocated automatically.
380 380
    ///The destructor deallocates this automatically allocated map,
381 381
    ///of course.
382 382
    ///\return <tt> (*this) </tt>
383 383
    Bfs &processedMap(ProcessedMap &m)
384 384
    {
385 385
      if(local_processed) {
386 386
        delete _processed;
387 387
        local_processed=false;
388 388
      }
389 389
      _processed = &m;
390 390
      return *this;
391 391
    }
392 392

	
393 393
    ///Sets the map that stores the distances of the nodes.
394 394

	
395 395
    ///Sets the map that stores the distances of the nodes calculated by
396 396
    ///the algorithm.
397 397
    ///If you don't use this function before calling \ref run(Node) "run()"
398 398
    ///or \ref init(), an instance will be allocated automatically.
399 399
    ///The destructor deallocates this automatically allocated map,
400 400
    ///of course.
401 401
    ///\return <tt> (*this) </tt>
402 402
    Bfs &distMap(DistMap &m)
403 403
    {
404 404
      if(local_dist) {
405 405
        delete _dist;
406 406
        local_dist=false;
407 407
      }
408 408
      _dist = &m;
409 409
      return *this;
410 410
    }
411 411

	
412 412
  public:
413 413

	
414 414
    ///\name Execution Control
415 415
    ///The simplest way to execute the BFS algorithm is to use one of the
416 416
    ///member functions called \ref run(Node) "run()".\n
417
    ///If you need more control on the execution, first you have to call
418
    ///\ref init(), then you can add several source nodes with
417
    ///If you need better control on the execution, you have to call
418
    ///\ref init() first, then you can add several source nodes with
419 419
    ///\ref addSource(). Finally the actual path computation can be
420 420
    ///performed with one of the \ref start() functions.
421 421

	
422 422
    ///@{
423 423

	
424 424
    ///\brief Initializes the internal data structures.
425 425
    ///
426 426
    ///Initializes the internal data structures.
427 427
    void init()
428 428
    {
429 429
      create_maps();
430 430
      _queue.resize(countNodes(*G));
431 431
      _queue_head=_queue_tail=0;
432 432
      _curr_dist=1;
433 433
      for ( NodeIt u(*G) ; u!=INVALID ; ++u ) {
434 434
        _pred->set(u,INVALID);
435 435
        _reached->set(u,false);
436 436
        _processed->set(u,false);
437 437
      }
438 438
    }
439 439

	
440 440
    ///Adds a new source node.
441 441

	
442 442
    ///Adds a new source node to the set of nodes to be processed.
443 443
    ///
444 444
    void addSource(Node s)
445 445
    {
446 446
      if(!(*_reached)[s])
447 447
        {
448 448
          _reached->set(s,true);
449 449
          _pred->set(s,INVALID);
450 450
          _dist->set(s,0);
451 451
          _queue[_queue_head++]=s;
452 452
          _queue_next_dist=_queue_head;
453 453
        }
454 454
    }
455 455

	
456 456
    ///Processes the next node.
457 457

	
458 458
    ///Processes the next node.
459 459
    ///
460 460
    ///\return The processed node.
461 461
    ///
462 462
    ///\pre The queue must not be empty.
463 463
    Node processNextNode()
464 464
    {
465 465
      if(_queue_tail==_queue_next_dist) {
466 466
        _curr_dist++;
467 467
        _queue_next_dist=_queue_head;
468 468
      }
469 469
      Node n=_queue[_queue_tail++];
470 470
      _processed->set(n,true);
471 471
      Node m;
472 472
      for(OutArcIt e(*G,n);e!=INVALID;++e)
473 473
        if(!(*_reached)[m=G->target(e)]) {
474 474
          _queue[_queue_head++]=m;
475 475
          _reached->set(m,true);
476 476
          _pred->set(m,e);
477 477
          _dist->set(m,_curr_dist);
478 478
        }
479 479
      return n;
480 480
    }
481 481

	
482 482
    ///Processes the next node.
483 483

	
484 484
    ///Processes the next node and checks if the given target node
485 485
    ///is reached. If the target node is reachable from the processed
486 486
    ///node, then the \c reach parameter will be set to \c true.
487 487
    ///
488 488
    ///\param target The target node.
489 489
    ///\retval reach Indicates if the target node is reached.
490 490
    ///It should be initially \c false.
491 491
    ///
492 492
    ///\return The processed node.
493 493
    ///
494 494
    ///\pre The queue must not be empty.
495 495
    Node processNextNode(Node target, bool& reach)
496 496
    {
497 497
      if(_queue_tail==_queue_next_dist) {
498 498
        _curr_dist++;
499 499
        _queue_next_dist=_queue_head;
500 500
      }
501 501
      Node n=_queue[_queue_tail++];
502 502
      _processed->set(n,true);
503 503
      Node m;
504 504
      for(OutArcIt e(*G,n);e!=INVALID;++e)
505 505
        if(!(*_reached)[m=G->target(e)]) {
506 506
          _queue[_queue_head++]=m;
507 507
          _reached->set(m,true);
508 508
          _pred->set(m,e);
509 509
          _dist->set(m,_curr_dist);
510 510
          reach = reach || (target == m);
511 511
        }
512 512
      return n;
513 513
    }
514 514

	
... ...
@@ -1329,194 +1329,194 @@
1329 1329
  private:
1330 1330

	
1331 1331
    typedef typename Digraph::Node Node;
1332 1332
    typedef typename Digraph::NodeIt NodeIt;
1333 1333
    typedef typename Digraph::Arc Arc;
1334 1334
    typedef typename Digraph::OutArcIt OutArcIt;
1335 1335

	
1336 1336
    //Pointer to the underlying digraph.
1337 1337
    const Digraph *_digraph;
1338 1338
    //Pointer to the visitor object.
1339 1339
    Visitor *_visitor;
1340 1340
    //Pointer to the map of reached status of the nodes.
1341 1341
    ReachedMap *_reached;
1342 1342
    //Indicates if _reached is locally allocated (true) or not.
1343 1343
    bool local_reached;
1344 1344

	
1345 1345
    std::vector<typename Digraph::Node> _list;
1346 1346
    int _list_front, _list_back;
1347 1347

	
1348 1348
    //Creates the maps if necessary.
1349 1349
    void create_maps() {
1350 1350
      if(!_reached) {
1351 1351
        local_reached = true;
1352 1352
        _reached = Traits::createReachedMap(*_digraph);
1353 1353
      }
1354 1354
    }
1355 1355

	
1356 1356
  protected:
1357 1357

	
1358 1358
    BfsVisit() {}
1359 1359

	
1360 1360
  public:
1361 1361

	
1362 1362
    typedef BfsVisit Create;
1363 1363

	
1364 1364
    /// \name Named Template Parameters
1365 1365

	
1366 1366
    ///@{
1367 1367
    template <class T>
1368 1368
    struct SetReachedMapTraits : public Traits {
1369 1369
      typedef T ReachedMap;
1370 1370
      static ReachedMap *createReachedMap(const Digraph &digraph) {
1371 1371
        LEMON_ASSERT(false, "ReachedMap is not initialized");
1372 1372
        return 0; // ignore warnings
1373 1373
      }
1374 1374
    };
1375 1375
    /// \brief \ref named-templ-param "Named parameter" for setting
1376 1376
    /// ReachedMap type.
1377 1377
    ///
1378 1378
    /// \ref named-templ-param "Named parameter" for setting ReachedMap type.
1379 1379
    template <class T>
1380 1380
    struct SetReachedMap : public BfsVisit< Digraph, Visitor,
1381 1381
                                            SetReachedMapTraits<T> > {
1382 1382
      typedef BfsVisit< Digraph, Visitor, SetReachedMapTraits<T> > Create;
1383 1383
    };
1384 1384
    ///@}
1385 1385

	
1386 1386
  public:
1387 1387

	
1388 1388
    /// \brief Constructor.
1389 1389
    ///
1390 1390
    /// Constructor.
1391 1391
    ///
1392 1392
    /// \param digraph The digraph the algorithm runs on.
1393 1393
    /// \param visitor The visitor object of the algorithm.
1394 1394
    BfsVisit(const Digraph& digraph, Visitor& visitor)
1395 1395
      : _digraph(&digraph), _visitor(&visitor),
1396 1396
        _reached(0), local_reached(false) {}
1397 1397

	
1398 1398
    /// \brief Destructor.
1399 1399
    ~BfsVisit() {
1400 1400
      if(local_reached) delete _reached;
1401 1401
    }
1402 1402

	
1403 1403
    /// \brief Sets the map that indicates which nodes are reached.
1404 1404
    ///
1405 1405
    /// Sets the map that indicates which nodes are reached.
1406 1406
    /// If you don't use this function before calling \ref run(Node) "run()"
1407 1407
    /// or \ref init(), an instance will be allocated automatically.
1408 1408
    /// The destructor deallocates this automatically allocated map,
1409 1409
    /// of course.
1410 1410
    /// \return <tt> (*this) </tt>
1411 1411
    BfsVisit &reachedMap(ReachedMap &m) {
1412 1412
      if(local_reached) {
1413 1413
        delete _reached;
1414 1414
        local_reached = false;
1415 1415
      }
1416 1416
      _reached = &m;
1417 1417
      return *this;
1418 1418
    }
1419 1419

	
1420 1420
  public:
1421 1421

	
1422 1422
    /// \name Execution Control
1423 1423
    /// The simplest way to execute the BFS algorithm is to use one of the
1424 1424
    /// member functions called \ref run(Node) "run()".\n
1425
    /// If you need more control on the execution, first you have to call
1426
    /// \ref init(), then you can add several source nodes with
1425
    /// If you need better control on the execution, you have to call
1426
    /// \ref init() first, then you can add several source nodes with
1427 1427
    /// \ref addSource(). Finally the actual path computation can be
1428 1428
    /// performed with one of the \ref start() functions.
1429 1429

	
1430 1430
    /// @{
1431 1431

	
1432 1432
    /// \brief Initializes the internal data structures.
1433 1433
    ///
1434 1434
    /// Initializes the internal data structures.
1435 1435
    void init() {
1436 1436
      create_maps();
1437 1437
      _list.resize(countNodes(*_digraph));
1438 1438
      _list_front = _list_back = -1;
1439 1439
      for (NodeIt u(*_digraph) ; u != INVALID ; ++u) {
1440 1440
        _reached->set(u, false);
1441 1441
      }
1442 1442
    }
1443 1443

	
1444 1444
    /// \brief Adds a new source node.
1445 1445
    ///
1446 1446
    /// Adds a new source node to the set of nodes to be processed.
1447 1447
    void addSource(Node s) {
1448 1448
      if(!(*_reached)[s]) {
1449 1449
          _reached->set(s,true);
1450 1450
          _visitor->start(s);
1451 1451
          _visitor->reach(s);
1452 1452
          _list[++_list_back] = s;
1453 1453
        }
1454 1454
    }
1455 1455

	
1456 1456
    /// \brief Processes the next node.
1457 1457
    ///
1458 1458
    /// Processes the next node.
1459 1459
    ///
1460 1460
    /// \return The processed node.
1461 1461
    ///
1462 1462
    /// \pre The queue must not be empty.
1463 1463
    Node processNextNode() {
1464 1464
      Node n = _list[++_list_front];
1465 1465
      _visitor->process(n);
1466 1466
      Arc e;
1467 1467
      for (_digraph->firstOut(e, n); e != INVALID; _digraph->nextOut(e)) {
1468 1468
        Node m = _digraph->target(e);
1469 1469
        if (!(*_reached)[m]) {
1470 1470
          _visitor->discover(e);
1471 1471
          _visitor->reach(m);
1472 1472
          _reached->set(m, true);
1473 1473
          _list[++_list_back] = m;
1474 1474
        } else {
1475 1475
          _visitor->examine(e);
1476 1476
        }
1477 1477
      }
1478 1478
      return n;
1479 1479
    }
1480 1480

	
1481 1481
    /// \brief Processes the next node.
1482 1482
    ///
1483 1483
    /// Processes the next node and checks if the given target node
1484 1484
    /// is reached. If the target node is reachable from the processed
1485 1485
    /// node, then the \c reach parameter will be set to \c true.
1486 1486
    ///
1487 1487
    /// \param target The target node.
1488 1488
    /// \retval reach Indicates if the target node is reached.
1489 1489
    /// It should be initially \c false.
1490 1490
    ///
1491 1491
    /// \return The processed node.
1492 1492
    ///
1493 1493
    /// \pre The queue must not be empty.
1494 1494
    Node processNextNode(Node target, bool& reach) {
1495 1495
      Node n = _list[++_list_front];
1496 1496
      _visitor->process(n);
1497 1497
      Arc e;
1498 1498
      for (_digraph->firstOut(e, n); e != INVALID; _digraph->nextOut(e)) {
1499 1499
        Node m = _digraph->target(e);
1500 1500
        if (!(*_reached)[m]) {
1501 1501
          _visitor->discover(e);
1502 1502
          _visitor->reach(m);
1503 1503
          _reached->set(m, true);
1504 1504
          _list[++_list_back] = m;
1505 1505
          reach = reach || (target == m);
1506 1506
        } else {
1507 1507
          _visitor->examine(e);
1508 1508
        }
1509 1509
      }
1510 1510
      return n;
1511 1511
    }
1512 1512

	
1513 1513
    /// \brief Processes the next node.
1514 1514
    ///
1515 1515
    /// Processes the next node and checks if at least one of reached
1516 1516
    /// nodes has \c true value in the \c nm node map. If one node
1517 1517
    /// with \c true value is reachable from the processed node, then the
1518 1518
    /// \c rnode parameter will be set to the first of such nodes.
1519 1519
    ///
1520 1520
    /// \param nm A \c bool (or convertible) node map that indicates the
1521 1521
    /// possible targets.
1522 1522
    /// \retval rnode The reached target node.
Ignore white space 6 line context
1 1
/* -*- mode: C++; indent-tabs-mode: nil; -*-
2 2
 *
3 3
 * This file is a part of LEMON, a generic C++ optimization library.
4 4
 *
5 5
 * Copyright (C) 2003-2009
6 6
 * Egervary Jeno Kombinatorikus Optimalizalasi Kutatocsoport
7 7
 * (Egervary Research Group on Combinatorial Optimization, EGRES).
8 8
 *
9 9
 * Permission to use, modify and distribute this software is granted
10 10
 * provided that this copyright notice appears in all copies. For
11 11
 * precise terms see the accompanying LICENSE file.
12 12
 *
13 13
 * This software is provided "AS IS" with no warranty of any kind,
14 14
 * express or implied, and with no claim as to its suitability for any
15 15
 * purpose.
16 16
 *
17 17
 */
18 18

	
19 19
#ifndef LEMON_BIN_HEAP_H
20 20
#define LEMON_BIN_HEAP_H
21 21

	
22
///\ingroup auxdat
22
///\ingroup heaps
23 23
///\file
24
///\brief Binary Heap implementation.
24
///\brief Binary heap implementation.
25 25

	
26 26
#include <vector>
27 27
#include <utility>
28 28
#include <functional>
29 29

	
30 30
namespace lemon {
31 31

	
32
  ///\ingroup auxdat
32
  /// \ingroup heaps
33 33
  ///
34
  ///\brief A Binary Heap implementation.
34
  /// \brief Binary heap data structure.
35 35
  ///
36
  ///This class implements the \e binary \e heap data structure. 
37
  /// 
38
  ///A \e heap is a data structure for storing items with specified values
39
  ///called \e priorities in such a way that finding the item with minimum
40
  ///priority is efficient. \c Comp specifies the ordering of the priorities.
41
  ///In a heap one can change the priority of an item, add or erase an
42
  ///item, etc.
36
  /// This class implements the \e binary \e heap data structure.
37
  /// It fully conforms to the \ref concepts::Heap "heap concept".
43 38
  ///
44
  ///\tparam PR Type of the priority of the items.
45
  ///\tparam IM A read and writable item map with int values, used internally
46
  ///to handle the cross references.
47
  ///\tparam Comp A functor class for the ordering of the priorities.
48
  ///The default is \c std::less<PR>.
49
  ///
50
  ///\sa FibHeap
51
  ///\sa Dijkstra
52
  template <typename PR, typename IM, typename Comp = std::less<PR> >
39
  /// \tparam PR Type of the priorities of the items.
40
  /// \tparam IM A read-writable item map with \c int values, used
41
  /// internally to handle the cross references.
42
  /// \tparam CMP A functor class for comparing the priorities.
43
  /// The default is \c std::less<PR>.
44
#ifdef DOXYGEN
45
  template <typename PR, typename IM, typename CMP>
46
#else
47
  template <typename PR, typename IM, typename CMP = std::less<PR> >
48
#endif
53 49
  class BinHeap {
50
  public:
54 51

	
55
  public:
56
    ///\e
52
    /// Type of the item-int map.
57 53
    typedef IM ItemIntMap;
58
    ///\e
54
    /// Type of the priorities.
59 55
    typedef PR Prio;
60
    ///\e
56
    /// Type of the items stored in the heap.
61 57
    typedef typename ItemIntMap::Key Item;
62
    ///\e
58
    /// Type of the item-priority pairs.
63 59
    typedef std::pair<Item,Prio> Pair;
64
    ///\e
65
    typedef Comp Compare;
60
    /// Functor type for comparing the priorities.
61
    typedef CMP Compare;
66 62

	
67
    /// \brief Type to represent the items states.
63
    /// \brief Type to represent the states of the items.
68 64
    ///
69
    /// Each Item element have a state associated to it. It may be "in heap",
70
    /// "pre heap" or "post heap". The latter two are indifferent from the
65
    /// Each item has a state associated to it. It can be "in heap",
66
    /// "pre-heap" or "post-heap". The latter two are indifferent from the
71 67
    /// heap's point of view, but may be useful to the user.
72 68
    ///
73 69
    /// The item-int map must be initialized in such way that it assigns
74 70
    /// \c PRE_HEAP (<tt>-1</tt>) to any element to be put in the heap.
75 71
    enum State {
76 72
      IN_HEAP = 0,    ///< = 0.
77 73
      PRE_HEAP = -1,  ///< = -1.
78 74
      POST_HEAP = -2  ///< = -2.
79 75
    };
80 76

	
81 77
  private:
82 78
    std::vector<Pair> _data;
83 79
    Compare _comp;
84 80
    ItemIntMap &_iim;
85 81

	
86 82
  public:
87
    /// \brief The constructor.
83

	
84
    /// \brief Constructor.
88 85
    ///
89
    /// The constructor.
90
    /// \param map should be given to the constructor, since it is used
91
    /// internally to handle the cross references. The value of the map
92
    /// must be \c PRE_HEAP (<tt>-1</tt>) for every item.
86
    /// Constructor.
87
    /// \param map A map that assigns \c int values to the items.
88
    /// It is used internally to handle the cross references.
89
    /// The assigned value must be \c PRE_HEAP (<tt>-1</tt>) for each item.
93 90
    explicit BinHeap(ItemIntMap &map) : _iim(map) {}
94 91

	
95
    /// \brief The constructor.
92
    /// \brief Constructor.
96 93
    ///
97
    /// The constructor.
98
    /// \param map should be given to the constructor, since it is used
99
    /// internally to handle the cross references. The value of the map
100
    /// should be PRE_HEAP (-1) for each element.
101
    ///
102
    /// \param comp The comparator function object.
94
    /// Constructor.
95
    /// \param map A map that assigns \c int values to the items.
96
    /// It is used internally to handle the cross references.
97
    /// The assigned value must be \c PRE_HEAP (<tt>-1</tt>) for each item.
98
    /// \param comp The function object used for comparing the priorities.
103 99
    BinHeap(ItemIntMap &map, const Compare &comp)
104 100
      : _iim(map), _comp(comp) {}
105 101

	
106 102

	
107
    /// The number of items stored in the heap.
103
    /// \brief The number of items stored in the heap.
108 104
    ///
109
    /// \brief Returns the number of items stored in the heap.
105
    /// This function returns the number of items stored in the heap.
110 106
    int size() const { return _data.size(); }
111 107

	
112
    /// \brief Checks if the heap stores no items.
108
    /// \brief Check if the heap is empty.
113 109
    ///
114
    /// Returns \c true if and only if the heap stores no items.
110
    /// This function returns \c true if the heap is empty.
115 111
    bool empty() const { return _data.empty(); }
116 112

	
117
    /// \brief Make empty this heap.
113
    /// \brief Make the heap empty.
118 114
    ///
119
    /// Make empty this heap. It does not change the cross reference map.
120
    /// If you want to reuse what is not surely empty you should first clear
121
    /// the heap and after that you should set the cross reference map for
122
    /// each item to \c PRE_HEAP.
115
    /// This functon makes the heap empty.
116
    /// It does not change the cross reference map. If you want to reuse
117
    /// a heap that is not surely empty, you should first clear it and
118
    /// then you should set the cross reference map to \c PRE_HEAP
119
    /// for each item.
123 120
    void clear() {
124 121
      _data.clear();
125 122
    }
126 123

	
127 124
  private:
128 125
    static int parent(int i) { return (i-1)/2; }
129 126

	
130
    static int second_child(int i) { return 2*i+2; }
127
    static int secondChild(int i) { return 2*i+2; }
131 128
    bool less(const Pair &p1, const Pair &p2) const {
132 129
      return _comp(p1.second, p2.second);
133 130
    }
134 131

	
135
    int bubble_up(int hole, Pair p) {
132
    int bubbleUp(int hole, Pair p) {
136 133
      int par = parent(hole);
137 134
      while( hole>0 && less(p,_data[par]) ) {
138 135
        move(_data[par],hole);
139 136
        hole = par;
140 137
        par = parent(hole);
141 138
      }
142 139
      move(p, hole);
143 140
      return hole;
144 141
    }
145 142

	
146
    int bubble_down(int hole, Pair p, int length) {
147
      int child = second_child(hole);
143
    int bubbleDown(int hole, Pair p, int length) {
144
      int child = secondChild(hole);
148 145
      while(child < length) {
149 146
        if( less(_data[child-1], _data[child]) ) {
150 147
          --child;
151 148
        }
152 149
        if( !less(_data[child], p) )
153 150
          goto ok;
154 151
        move(_data[child], hole);
155 152
        hole = child;
156
        child = second_child(hole);
153
        child = secondChild(hole);
157 154
      }
158 155
      child--;
159 156
      if( child<length && less(_data[child], p) ) {
160 157
        move(_data[child], hole);
161 158
        hole=child;
162 159
      }
163 160
    ok:
164 161
      move(p, hole);
165 162
      return hole;
166 163
    }
167 164

	
168 165
    void move(const Pair &p, int i) {
169 166
      _data[i] = p;
170 167
      _iim.set(p.first, i);
171 168
    }
172 169

	
173 170
  public:
171

	
174 172
    /// \brief Insert a pair of item and priority into the heap.
175 173
    ///
176
    /// Adds \c p.first to the heap with priority \c p.second.
174
    /// This function inserts \c p.first to the heap with priority
175
    /// \c p.second.
177 176
    /// \param p The pair to insert.
177
    /// \pre \c p.first must not be stored in the heap.
178 178
    void push(const Pair &p) {
179 179
      int n = _data.size();
180 180
      _data.resize(n+1);
181
      bubble_up(n, p);
181
      bubbleUp(n, p);
182 182
    }
183 183

	
184
    /// \brief Insert an item into the heap with the given heap.
184
    /// \brief Insert an item into the heap with the given priority.
185 185
    ///
186
    /// Adds \c i to the heap with priority \c p.
186
    /// This function inserts the given item into the heap with the
187
    /// given priority.
187 188
    /// \param i The item to insert.
188 189
    /// \param p The priority of the item.
190
    /// \pre \e i must not be stored in the heap.
189 191
    void push(const Item &i, const Prio &p) { push(Pair(i,p)); }
190 192

	
191
    /// \brief Returns the item with minimum priority relative to \c Compare.
193
    /// \brief Return the item having minimum priority.
192 194
    ///
193
    /// This method returns the item with minimum priority relative to \c
194
    /// Compare.
195
    /// \pre The heap must be nonempty.
195
    /// This function returns the item having minimum priority.
196
    /// \pre The heap must be non-empty.
196 197
    Item top() const {
197 198
      return _data[0].first;
198 199
    }
199 200

	
200
    /// \brief Returns the minimum priority relative to \c Compare.
201
    /// \brief The minimum priority.
201 202
    ///
202
    /// It returns the minimum priority relative to \c Compare.
203
    /// \pre The heap must be nonempty.
203
    /// This function returns the minimum priority.
204
    /// \pre The heap must be non-empty.
204 205
    Prio prio() const {
205 206
      return _data[0].second;
206 207
    }
207 208

	
208
    /// \brief Deletes the item with minimum priority relative to \c Compare.
209
    /// \brief Remove the item having minimum priority.
209 210
    ///
210
    /// This method deletes the item with minimum priority relative to \c
211
    /// Compare from the heap.
211
    /// This function removes the item having minimum priority.
212 212
    /// \pre The heap must be non-empty.
213 213
    void pop() {
214 214
      int n = _data.size()-1;
215 215
      _iim.set(_data[0].first, POST_HEAP);
216 216
      if (n > 0) {
217
        bubble_down(0, _data[n], n);
217
        bubbleDown(0, _data[n], n);
218 218
      }
219 219
      _data.pop_back();
220 220
    }
221 221

	
222
    /// \brief Deletes \c i from the heap.
222
    /// \brief Remove the given item from the heap.
223 223
    ///
224
    /// This method deletes item \c i from the heap.
225
    /// \param i The item to erase.
226
    /// \pre The item should be in the heap.
224
    /// This function removes the given item from the heap if it is
225
    /// already stored.
226
    /// \param i The item to delete.
227
    /// \pre \e i must be in the heap.
227 228
    void erase(const Item &i) {
228 229
      int h = _iim[i];
229 230
      int n = _data.size()-1;
230 231
      _iim.set(_data[h].first, POST_HEAP);
231 232
      if( h < n ) {
232
        if ( bubble_up(h, _data[n]) == h) {
233
          bubble_down(h, _data[n], n);
233
        if ( bubbleUp(h, _data[n]) == h) {
234
          bubbleDown(h, _data[n], n);
234 235
        }
235 236
      }
236 237
      _data.pop_back();
237 238
    }
238 239

	
239

	
240
    /// \brief Returns the priority of \c i.
240
    /// \brief The priority of the given item.
241 241
    ///
242
    /// This function returns the priority of item \c i.
242
    /// This function returns the priority of the given item.
243 243
    /// \param i The item.
244
    /// \pre \c i must be in the heap.
244
    /// \pre \e i must be in the heap.
245 245
    Prio operator[](const Item &i) const {
246 246
      int idx = _iim[i];
247 247
      return _data[idx].second;
248 248
    }
249 249

	
250
    /// \brief \c i gets to the heap with priority \c p independently
251
    /// if \c i was already there.
250
    /// \brief Set the priority of an item or insert it, if it is
251
    /// not stored in the heap.
252 252
    ///
253
    /// This method calls \ref push(\c i, \c p) if \c i is not stored
254
    /// in the heap and sets the priority of \c i to \c p otherwise.
253
    /// This method sets the priority of the given item if it is
254
    /// already stored in the heap. Otherwise it inserts the given
255
    /// item into the heap with the given priority.
255 256
    /// \param i The item.
256 257
    /// \param p The priority.
257 258
    void set(const Item &i, const Prio &p) {
258 259
      int idx = _iim[i];
259 260
      if( idx < 0 ) {
260 261
        push(i,p);
261 262
      }
262 263
      else if( _comp(p, _data[idx].second) ) {
263
        bubble_up(idx, Pair(i,p));
264
        bubbleUp(idx, Pair(i,p));
264 265
      }
265 266
      else {
266
        bubble_down(idx, Pair(i,p), _data.size());
267
        bubbleDown(idx, Pair(i,p), _data.size());
267 268
      }
268 269
    }
269 270

	
270
    /// \brief Decreases the priority of \c i to \c p.
271
    /// \brief Decrease the priority of an item to the given value.
271 272
    ///
272
    /// This method decreases the priority of item \c i to \c p.
273
    /// This function decreases the priority of an item to the given value.
273 274
    /// \param i The item.
274 275
    /// \param p The priority.
275
    /// \pre \c i must be stored in the heap with priority at least \c
276
    /// p relative to \c Compare.
276
    /// \pre \e i must be stored in the heap with priority at least \e p.
277 277
    void decrease(const Item &i, const Prio &p) {
278 278
      int idx = _iim[i];
279
      bubble_up(idx, Pair(i,p));
279
      bubbleUp(idx, Pair(i,p));
280 280
    }
281 281

	
282
    /// \brief Increases the priority of \c i to \c p.
282
    /// \brief Increase the priority of an item to the given value.
283 283
    ///
284
    /// This method sets the priority of item \c i to \c p.
284
    /// This function increases the priority of an item to the given value.
285 285
    /// \param i The item.
286 286
    /// \param p The priority.
287
    /// \pre \c i must be stored in the heap with priority at most \c
288
    /// p relative to \c Compare.
287
    /// \pre \e i must be stored in the heap with priority at most \e p.
289 288
    void increase(const Item &i, const Prio &p) {
290 289
      int idx = _iim[i];
291
      bubble_down(idx, Pair(i,p), _data.size());
290
      bubbleDown(idx, Pair(i,p), _data.size());
292 291
    }
293 292

	
294
    /// \brief Returns if \c item is in, has already been in, or has
295
    /// never been in the heap.
293
    /// \brief Return the state of an item.
296 294
    ///
297
    /// This method returns PRE_HEAP if \c item has never been in the
298
    /// heap, IN_HEAP if it is in the heap at the moment, and POST_HEAP
299
    /// otherwise. In the latter case it is possible that \c item will
300
    /// get back to the heap again.
295
    /// This method returns \c PRE_HEAP if the given item has never
296
    /// been in the heap, \c IN_HEAP if it is in the heap at the moment,
297
    /// and \c POST_HEAP otherwise.
298
    /// In the latter case it is possible that the item will get back
299
    /// to the heap again.
301 300
    /// \param i The item.
302 301
    State state(const Item &i) const {
303 302
      int s = _iim[i];
304 303
      if( s>=0 )
305 304
        s=0;
306 305
      return State(s);
307 306
    }
308 307

	
309
    /// \brief Sets the state of the \c item in the heap.
308
    /// \brief Set the state of an item in the heap.
310 309
    ///
311
    /// Sets the state of the \c item in the heap. It can be used to
312
    /// manually clear the heap when it is important to achive the
313
    /// better time complexity.
310
    /// This function sets the state of the given item in the heap.
311
    /// It can be used to manually clear the heap when it is important
312
    /// to achive better time complexity.
314 313
    /// \param i The item.
315 314
    /// \param st The state. It should not be \c IN_HEAP.
316 315
    void state(const Item& i, State st) {
317 316
      switch (st) {
318 317
      case POST_HEAP:
319 318
      case PRE_HEAP:
320 319
        if (state(i) == IN_HEAP) {
321 320
          erase(i);
322 321
        }
323 322
        _iim[i] = st;
324 323
        break;
325 324
      case IN_HEAP:
326 325
        break;
327 326
      }
328 327
    }
329 328

	
330
    /// \brief Replaces an item in the heap.
329
    /// \brief Replace an item in the heap.
331 330
    ///
332
    /// The \c i item is replaced with \c j item. The \c i item should
333
    /// be in the heap, while the \c j should be out of the heap. The
334
    /// \c i item will out of the heap and \c j will be in the heap
335
    /// with the same prioriority as prevoiusly the \c i item.
331
    /// This function replaces item \c i with item \c j.
332
    /// Item \c i must be in the heap, while \c j must be out of the heap.
333
    /// After calling this method, item \c i will be out of the
334
    /// heap and \c j will be in the heap with the same prioriority
335
    /// as item \c i had before.
336 336
    void replace(const Item& i, const Item& j) {
337 337
      int idx = _iim[i];
338 338
      _iim.set(i, _iim[j]);
339 339
      _iim.set(j, idx);
340 340
      _data[idx].first = j;
341 341
    }
342 342

	
343 343
  }; // class BinHeap
344 344

	
345 345
} // namespace lemon
346 346

	
347 347
#endif // LEMON_BIN_HEAP_H
Ignore white space 6 line context
... ...
@@ -444,182 +444,182 @@
444 444

	
445 445
    class EdgeIt : public Parent::Edge { 
446 446
      const Graph* graph;
447 447
    public:
448 448

	
449 449
      EdgeIt() { }
450 450

	
451 451
      EdgeIt(Invalid i) : Edge(i) { }
452 452

	
453 453
      explicit EdgeIt(const Graph& _graph) : graph(&_graph) {
454 454
	_graph.first(static_cast<Edge&>(*this));
455 455
      }
456 456

	
457 457
      EdgeIt(const Graph& _graph, const Edge& e) : 
458 458
	Edge(e), graph(&_graph) { }
459 459

	
460 460
      EdgeIt& operator++() { 
461 461
	graph->next(*this);
462 462
	return *this; 
463 463
      }
464 464

	
465 465
    };
466 466

	
467 467
    class IncEdgeIt : public Parent::Edge {
468 468
      friend class EdgeSetExtender;
469 469
      const Graph* graph;
470 470
      bool direction;
471 471
    public:
472 472

	
473 473
      IncEdgeIt() { }
474 474

	
475 475
      IncEdgeIt(Invalid i) : Edge(i), direction(false) { }
476 476

	
477 477
      IncEdgeIt(const Graph& _graph, const Node &n) : graph(&_graph) {
478 478
	_graph.firstInc(*this, direction, n);
479 479
      }
480 480

	
481 481
      IncEdgeIt(const Graph& _graph, const Edge &ue, const Node &n)
482 482
	: graph(&_graph), Edge(ue) {
483 483
	direction = (_graph.source(ue) == n);
484 484
      }
485 485

	
486 486
      IncEdgeIt& operator++() {
487 487
	graph->nextInc(*this, direction);
488 488
	return *this; 
489 489
      }
490 490
    };
491 491

	
492 492
    // \brief Base node of the iterator
493 493
    //
494 494
    // Returns the base node (ie. the source in this case) of the iterator
495 495
    Node baseNode(const OutArcIt &e) const {
496 496
      return Parent::source(static_cast<const Arc&>(e));
497 497
    }
498 498
    // \brief Running node of the iterator
499 499
    //
500 500
    // Returns the running node (ie. the target in this case) of the
501 501
    // iterator
502 502
    Node runningNode(const OutArcIt &e) const {
503 503
      return Parent::target(static_cast<const Arc&>(e));
504 504
    }
505 505

	
506 506
    // \brief Base node of the iterator
507 507
    //
508 508
    // Returns the base node (ie. the target in this case) of the iterator
509 509
    Node baseNode(const InArcIt &e) const {
510 510
      return Parent::target(static_cast<const Arc&>(e));
511 511
    }
512 512
    // \brief Running node of the iterator
513 513
    //
514 514
    // Returns the running node (ie. the source in this case) of the
515 515
    // iterator
516 516
    Node runningNode(const InArcIt &e) const {
517 517
      return Parent::source(static_cast<const Arc&>(e));
518 518
    }
519 519

	
520 520
    // Base node of the iterator
521 521
    //
522 522
    // Returns the base node of the iterator
523 523
    Node baseNode(const IncEdgeIt &e) const {
524 524
      return e.direction ? u(e) : v(e);
525 525
    }
526 526
    // Running node of the iterator
527 527
    //
528 528
    // Returns the running node of the iterator
529 529
    Node runningNode(const IncEdgeIt &e) const {
530 530
      return e.direction ? v(e) : u(e);
531 531
    }
532 532

	
533 533

	
534 534
    template <typename _Value>
535 535
    class ArcMap 
536 536
      : public MapExtender<DefaultMap<Graph, Arc, _Value> > {
537 537
      typedef MapExtender<DefaultMap<Graph, Arc, _Value> > Parent;
538 538

	
539 539
    public:
540
      ArcMap(const Graph& _g) 
540
      explicit ArcMap(const Graph& _g) 
541 541
	: Parent(_g) {}
542 542
      ArcMap(const Graph& _g, const _Value& _v) 
543 543
	: Parent(_g, _v) {}
544 544

	
545 545
      ArcMap& operator=(const ArcMap& cmap) {
546 546
	return operator=<ArcMap>(cmap);
547 547
      }
548 548

	
549 549
      template <typename CMap>
550 550
      ArcMap& operator=(const CMap& cmap) {
551 551
        Parent::operator=(cmap);
552 552
	return *this;
553 553
      }
554 554

	
555 555
    };
556 556

	
557 557

	
558 558
    template <typename _Value>
559 559
    class EdgeMap 
560 560
      : public MapExtender<DefaultMap<Graph, Edge, _Value> > {
561 561
      typedef MapExtender<DefaultMap<Graph, Edge, _Value> > Parent;
562 562

	
563 563
    public:
564
      EdgeMap(const Graph& _g) 
564
      explicit EdgeMap(const Graph& _g) 
565 565
	: Parent(_g) {}
566 566

	
567 567
      EdgeMap(const Graph& _g, const _Value& _v) 
568 568
	: Parent(_g, _v) {}
569 569

	
570 570
      EdgeMap& operator=(const EdgeMap& cmap) {
571 571
	return operator=<EdgeMap>(cmap);
572 572
      }
573 573

	
574 574
      template <typename CMap>
575 575
      EdgeMap& operator=(const CMap& cmap) {
576 576
        Parent::operator=(cmap);
577 577
	return *this;
578 578
      }
579 579

	
580 580
    };
581 581

	
582 582

	
583 583
    // Alteration extension
584 584

	
585 585
    Edge addEdge(const Node& from, const Node& to) {
586 586
      Edge edge = Parent::addEdge(from, to);
587 587
      notifier(Edge()).add(edge);
588 588
      std::vector<Arc> arcs;
589 589
      arcs.push_back(Parent::direct(edge, true));
590 590
      arcs.push_back(Parent::direct(edge, false));
591 591
      notifier(Arc()).add(arcs);
592 592
      return edge;
593 593
    }
594 594
    
595 595
    void clear() {
596 596
      notifier(Arc()).clear();
597 597
      notifier(Edge()).clear();
598 598
      Parent::clear();
599 599
    }
600 600

	
601 601
    void erase(const Edge& edge) {
602 602
      std::vector<Arc> arcs;
603 603
      arcs.push_back(Parent::direct(edge, true));
604 604
      arcs.push_back(Parent::direct(edge, false));
605 605
      notifier(Arc()).erase(arcs);
606 606
      notifier(Edge()).erase(edge);
607 607
      Parent::erase(edge);
608 608
    }
609 609

	
610 610

	
611 611
    EdgeSetExtender() {
612 612
      arc_notifier.setContainer(*this);
613 613
      edge_notifier.setContainer(*this);
614 614
    }
615 615

	
616 616
    ~EdgeSetExtender() {
617 617
      edge_notifier.clear();
618 618
      arc_notifier.clear();
619 619
    }
620 620
    
621 621
  };
622 622

	
623 623
}
624 624

	
625 625
#endif
Ignore white space 6 line context
... ...
@@ -511,241 +511,241 @@
511 511
    class EdgeIt : public Parent::Edge {
512 512
      const Graph* _graph;
513 513
    public:
514 514

	
515 515
      EdgeIt() { }
516 516

	
517 517
      EdgeIt(Invalid i) : Edge(i) { }
518 518

	
519 519
      explicit EdgeIt(const Graph& graph) : _graph(&graph) {
520 520
        _graph->first(static_cast<Edge&>(*this));
521 521
      }
522 522

	
523 523
      EdgeIt(const Graph& graph, const Edge& edge) :
524 524
        Edge(edge), _graph(&graph) { }
525 525

	
526 526
      EdgeIt& operator++() {
527 527
        _graph->next(*this);
528 528
        return *this;
529 529
      }
530 530

	
531 531
    };
532 532

	
533 533
    class IncEdgeIt : public Parent::Edge {
534 534
      friend class GraphExtender;
535 535
      const Graph* _graph;
536 536
      bool _direction;
537 537
    public:
538 538

	
539 539
      IncEdgeIt() { }
540 540

	
541 541
      IncEdgeIt(Invalid i) : Edge(i), _direction(false) { }
542 542

	
543 543
      IncEdgeIt(const Graph& graph, const Node &node) : _graph(&graph) {
544 544
        _graph->firstInc(*this, _direction, node);
545 545
      }
546 546

	
547 547
      IncEdgeIt(const Graph& graph, const Edge &edge, const Node &node)
548 548
        : _graph(&graph), Edge(edge) {
549 549
        _direction = (_graph->source(edge) == node);
550 550
      }
551 551

	
552 552
      IncEdgeIt& operator++() {
553 553
        _graph->nextInc(*this, _direction);
554 554
        return *this;
555 555
      }
556 556
    };
557 557

	
558 558
    // \brief Base node of the iterator
559 559
    //
560 560
    // Returns the base node (ie. the source in this case) of the iterator
561 561
    Node baseNode(const OutArcIt &arc) const {
562 562
      return Parent::source(static_cast<const Arc&>(arc));
563 563
    }
564 564
    // \brief Running node of the iterator
565 565
    //
566 566
    // Returns the running node (ie. the target in this case) of the
567 567
    // iterator
568 568
    Node runningNode(const OutArcIt &arc) const {
569 569
      return Parent::target(static_cast<const Arc&>(arc));
570 570
    }
571 571

	
572 572
    // \brief Base node of the iterator
573 573
    //
574 574
    // Returns the base node (ie. the target in this case) of the iterator
575 575
    Node baseNode(const InArcIt &arc) const {
576 576
      return Parent::target(static_cast<const Arc&>(arc));
577 577
    }
578 578
    // \brief Running node of the iterator
579 579
    //
580 580
    // Returns the running node (ie. the source in this case) of the
581 581
    // iterator
582 582
    Node runningNode(const InArcIt &arc) const {
583 583
      return Parent::source(static_cast<const Arc&>(arc));
584 584
    }
585 585

	
586 586
    // Base node of the iterator
587 587
    //
588 588
    // Returns the base node of the iterator
589 589
    Node baseNode(const IncEdgeIt &edge) const {
590 590
      return edge._direction ? u(edge) : v(edge);
591 591
    }
592 592
    // Running node of the iterator
593 593
    //
594 594
    // Returns the running node of the iterator
595 595
    Node runningNode(const IncEdgeIt &edge) const {
596 596
      return edge._direction ? v(edge) : u(edge);
597 597
    }
598 598

	
599 599
    // Mappable extension
600 600

	
601 601
    template <typename _Value>
602 602
    class NodeMap
603 603
      : public MapExtender<DefaultMap<Graph, Node, _Value> > {
604 604
      typedef MapExtender<DefaultMap<Graph, Node, _Value> > Parent;
605 605

	
606 606
    public:
607
      NodeMap(const Graph& graph)
607
      explicit NodeMap(const Graph& graph)
608 608
        : Parent(graph) {}
609 609
      NodeMap(const Graph& graph, const _Value& value)
610 610
        : Parent(graph, value) {}
611 611

	
612 612
    private:
613 613
      NodeMap& operator=(const NodeMap& cmap) {
614 614
        return operator=<NodeMap>(cmap);
615 615
      }
616 616

	
617 617
      template <typename CMap>
618 618
      NodeMap& operator=(const CMap& cmap) {
619 619
        Parent::operator=(cmap);
620 620
        return *this;
621 621
      }
622 622

	
623 623
    };
624 624

	
625 625
    template <typename _Value>
626 626
    class ArcMap
627 627
      : public MapExtender<DefaultMap<Graph, Arc, _Value> > {
628 628
      typedef MapExtender<DefaultMap<Graph, Arc, _Value> > Parent;
629 629

	
630 630
    public:
631
      ArcMap(const Graph& graph)
631
      explicit ArcMap(const Graph& graph)
632 632
        : Parent(graph) {}
633 633
      ArcMap(const Graph& graph, const _Value& value)
634 634
        : Parent(graph, value) {}
635 635

	
636 636
    private:
637 637
      ArcMap& operator=(const ArcMap& cmap) {
638 638
        return operator=<ArcMap>(cmap);
639 639
      }
640 640

	
641 641
      template <typename CMap>
642 642
      ArcMap& operator=(const CMap& cmap) {
643 643
        Parent::operator=(cmap);
644 644
        return *this;
645 645
      }
646 646
    };
647 647

	
648 648

	
649 649
    template <typename _Value>
650 650
    class EdgeMap
651 651
      : public MapExtender<DefaultMap<Graph, Edge, _Value> > {
652 652
      typedef MapExtender<DefaultMap<Graph, Edge, _Value> > Parent;
653 653

	
654 654
    public:
655
      EdgeMap(const Graph& graph)
655
      explicit EdgeMap(const Graph& graph)
656 656
        : Parent(graph) {}
657 657

	
658 658
      EdgeMap(const Graph& graph, const _Value& value)
659 659
        : Parent(graph, value) {}
660 660

	
661 661
    private:
662 662
      EdgeMap& operator=(const EdgeMap& cmap) {
663 663
        return operator=<EdgeMap>(cmap);
664 664
      }
665 665

	
666 666
      template <typename CMap>
667 667
      EdgeMap& operator=(const CMap& cmap) {
668 668
        Parent::operator=(cmap);
669 669
        return *this;
670 670
      }
671 671

	
672 672
    };
673 673

	
674 674
    // Alteration extension
675 675

	
676 676
    Node addNode() {
677 677
      Node node = Parent::addNode();
678 678
      notifier(Node()).add(node);
679 679
      return node;
680 680
    }
681 681

	
682 682
    Edge addEdge(const Node& from, const Node& to) {
683 683
      Edge edge = Parent::addEdge(from, to);
684 684
      notifier(Edge()).add(edge);
685 685
      std::vector<Arc> ev;
686 686
      ev.push_back(Parent::direct(edge, true));
687 687
      ev.push_back(Parent::direct(edge, false));
688 688
      notifier(Arc()).add(ev);
689 689
      return edge;
690 690
    }
691 691

	
692 692
    void clear() {
693 693
      notifier(Arc()).clear();
694 694
      notifier(Edge()).clear();
695 695
      notifier(Node()).clear();
696 696
      Parent::clear();
697 697
    }
698 698

	
699 699
    template <typename Graph, typename NodeRefMap, typename EdgeRefMap>
700 700
    void build(const Graph& graph, NodeRefMap& nodeRef,
701 701
               EdgeRefMap& edgeRef) {
702 702
      Parent::build(graph, nodeRef, edgeRef);
703 703
      notifier(Node()).build();
704 704
      notifier(Edge()).build();
705 705
      notifier(Arc()).build();
706 706
    }
707 707

	
708 708
    void erase(const Node& node) {
709 709
      Arc arc;
710 710
      Parent::firstOut(arc, node);
711 711
      while (arc != INVALID ) {
712 712
        erase(arc);
713 713
        Parent::firstOut(arc, node);
714 714
      }
715 715

	
716 716
      Parent::firstIn(arc, node);
717 717
      while (arc != INVALID ) {
718 718
        erase(arc);
719 719
        Parent::firstIn(arc, node);
720 720
      }
721 721

	
722 722
      notifier(Node()).erase(node);
723 723
      Parent::erase(node);
724 724
    }
725 725

	
726 726
    void erase(const Edge& edge) {
727 727
      std::vector<Arc> av;
728 728
      av.push_back(Parent::direct(edge, true));
729 729
      av.push_back(Parent::direct(edge, false));
730 730
      notifier(Arc()).erase(av);
731 731
      notifier(Edge()).erase(edge);
732 732
      Parent::erase(edge);
733 733
    }
734 734

	
735 735
    GraphExtender() {
736 736
      node_notifier.setContainer(*this);
737 737
      arc_notifier.setContainer(*this);
738 738
      edge_notifier.setContainer(*this);
739 739
    }
740 740

	
741 741
    ~GraphExtender() {
742 742
      edge_notifier.clear();
743 743
      arc_notifier.clear();
744 744
      node_notifier.clear();
745 745
    }
746 746

	
747 747
  };
748 748

	
749 749
}
750 750

	
751 751
#endif
Ignore white space 6 line context
1 1
/* -*- mode: C++; indent-tabs-mode: nil; -*-
2 2
 *
3 3
 * This file is a part of LEMON, a generic C++ optimization library.
4 4
 *
5 5
 * Copyright (C) 2003-2009
6 6
 * Egervary Jeno Kombinatorikus Optimalizalasi Kutatocsoport
7 7
 * (Egervary Research Group on Combinatorial Optimization, EGRES).
8 8
 *
9 9
 * Permission to use, modify and distribute this software is granted
10 10
 * provided that this copyright notice appears in all copies. For
11 11
 * precise terms see the accompanying LICENSE file.
12 12
 *
13 13
 * This software is provided "AS IS" with no warranty of any kind,
14 14
 * express or implied, and with no claim as to its suitability for any
15 15
 * purpose.
16 16
 *
17 17
 */
18 18

	
19 19
#ifndef LEMON_CIRCULATION_H
20 20
#define LEMON_CIRCULATION_H
21 21

	
22 22
#include <lemon/tolerance.h>
23 23
#include <lemon/elevator.h>
24 24
#include <limits>
25 25

	
26 26
///\ingroup max_flow
27 27
///\file
28 28
///\brief Push-relabel algorithm for finding a feasible circulation.
29 29
///
30 30
namespace lemon {
31 31

	
32 32
  /// \brief Default traits class of Circulation class.
33 33
  ///
34 34
  /// Default traits class of Circulation class.
35 35
  ///
36 36
  /// \tparam GR Type of the digraph the algorithm runs on.
37 37
  /// \tparam LM The type of the lower bound map.
38 38
  /// \tparam UM The type of the upper bound (capacity) map.
39 39
  /// \tparam SM The type of the supply map.
40 40
  template <typename GR, typename LM,
41 41
            typename UM, typename SM>
42 42
  struct CirculationDefaultTraits {
43 43

	
44 44
    /// \brief The type of the digraph the algorithm runs on.
45 45
    typedef GR Digraph;
46 46

	
47 47
    /// \brief The type of the lower bound map.
48 48
    ///
49 49
    /// The type of the map that stores the lower bounds on the arcs.
50 50
    /// It must conform to the \ref concepts::ReadMap "ReadMap" concept.
51 51
    typedef LM LowerMap;
52 52

	
53 53
    /// \brief The type of the upper bound (capacity) map.
54 54
    ///
55 55
    /// The type of the map that stores the upper bounds (capacities)
56 56
    /// on the arcs.
57 57
    /// It must conform to the \ref concepts::ReadMap "ReadMap" concept.
58 58
    typedef UM UpperMap;
59 59

	
60 60
    /// \brief The type of supply map.
61 61
    ///
62 62
    /// The type of the map that stores the signed supply values of the 
63 63
    /// nodes. 
64 64
    /// It must conform to the \ref concepts::ReadMap "ReadMap" concept.
65 65
    typedef SM SupplyMap;
66 66

	
67 67
    /// \brief The type of the flow and supply values.
68 68
    typedef typename SupplyMap::Value Value;
69 69

	
70 70
    /// \brief The type of the map that stores the flow values.
71 71
    ///
72 72
    /// The type of the map that stores the flow values.
73 73
    /// It must conform to the \ref concepts::ReadWriteMap "ReadWriteMap"
74 74
    /// concept.
75
#ifdef DOXYGEN
76
    typedef GR::ArcMap<Value> FlowMap;
77
#else
75 78
    typedef typename Digraph::template ArcMap<Value> FlowMap;
79
#endif
76 80

	
77 81
    /// \brief Instantiates a FlowMap.
78 82
    ///
79 83
    /// This function instantiates a \ref FlowMap.
80 84
    /// \param digraph The digraph for which we would like to define
81 85
    /// the flow map.
82 86
    static FlowMap* createFlowMap(const Digraph& digraph) {
83 87
      return new FlowMap(digraph);
84 88
    }
85 89

	
86 90
    /// \brief The elevator type used by the algorithm.
87 91
    ///
88 92
    /// The elevator type used by the algorithm.
89 93
    ///
90
    /// \sa Elevator
91
    /// \sa LinkedElevator
94
    /// \sa Elevator, LinkedElevator
95
#ifdef DOXYGEN
96
    typedef lemon::Elevator<GR, GR::Node> Elevator;
97
#else
92 98
    typedef lemon::Elevator<Digraph, typename Digraph::Node> Elevator;
99
#endif
93 100

	
94 101
    /// \brief Instantiates an Elevator.
95 102
    ///
96 103
    /// This function instantiates an \ref Elevator.
97 104
    /// \param digraph The digraph for which we would like to define
98 105
    /// the elevator.
99 106
    /// \param max_level The maximum level of the elevator.
100 107
    static Elevator* createElevator(const Digraph& digraph, int max_level) {
101 108
      return new Elevator(digraph, max_level);
102 109
    }
103 110

	
104 111
    /// \brief The tolerance used by the algorithm
105 112
    ///
106 113
    /// The tolerance used by the algorithm to handle inexact computation.
107 114
    typedef lemon::Tolerance<Value> Tolerance;
108 115

	
109 116
  };
110 117

	
111 118
  /**
112 119
     \brief Push-relabel algorithm for the network circulation problem.
113 120

	
114 121
     \ingroup max_flow
115 122
     This class implements a push-relabel algorithm for the \e network
116 123
     \e circulation problem.
117 124
     It is to find a feasible circulation when lower and upper bounds
118 125
     are given for the flow values on the arcs and lower bounds are
119 126
     given for the difference between the outgoing and incoming flow
120 127
     at the nodes.
121 128

	
122 129
     The exact formulation of this problem is the following.
123 130
     Let \f$G=(V,A)\f$ be a digraph, \f$lower: A\rightarrow\mathbf{R}\f$
124 131
     \f$upper: A\rightarrow\mathbf{R}\cup\{\infty\}\f$ denote the lower and
125 132
     upper bounds on the arcs, for which \f$lower(uv) \leq upper(uv)\f$
126 133
     holds for all \f$uv\in A\f$, and \f$sup: V\rightarrow\mathbf{R}\f$
127 134
     denotes the signed supply values of the nodes.
128 135
     If \f$sup(u)>0\f$, then \f$u\f$ is a supply node with \f$sup(u)\f$
129 136
     supply, if \f$sup(u)<0\f$, then \f$u\f$ is a demand node with
130 137
     \f$-sup(u)\f$ demand.
131 138
     A feasible circulation is an \f$f: A\rightarrow\mathbf{R}\f$
132 139
     solution of the following problem.
133 140

	
134 141
     \f[ \sum_{uv\in A} f(uv) - \sum_{vu\in A} f(vu)
135 142
     \geq sup(u) \quad \forall u\in V, \f]
136 143
     \f[ lower(uv) \leq f(uv) \leq upper(uv) \quad \forall uv\in A. \f]
137 144
     
138 145
     The sum of the supply values, i.e. \f$\sum_{u\in V} sup(u)\f$ must be
139 146
     zero or negative in order to have a feasible solution (since the sum
140 147
     of the expressions on the left-hand side of the inequalities is zero).
141 148
     It means that the total demand must be greater or equal to the total
142 149
     supply and all the supplies have to be carried out from the supply nodes,
143 150
     but there could be demands that are not satisfied.
144 151
     If \f$\sum_{u\in V} sup(u)\f$ is zero, then all the supply/demand
145 152
     constraints have to be satisfied with equality, i.e. all demands
146 153
     have to be satisfied and all supplies have to be used.
147 154
     
148 155
     If you need the opposite inequalities in the supply/demand constraints
149 156
     (i.e. the total demand is less than the total supply and all the demands
150 157
     have to be satisfied while there could be supplies that are not used),
151 158
     then you could easily transform the problem to the above form by reversing
152 159
     the direction of the arcs and taking the negative of the supply values
153 160
     (e.g. using \ref ReverseDigraph and \ref NegMap adaptors).
154 161

	
155 162
     This algorithm either calculates a feasible circulation, or provides
156 163
     a \ref barrier() "barrier", which prooves that a feasible soultion
157 164
     cannot exist.
158 165

	
159 166
     Note that this algorithm also provides a feasible solution for the
160 167
     \ref min_cost_flow "minimum cost flow problem".
161 168

	
162 169
     \tparam GR The type of the digraph the algorithm runs on.
163 170
     \tparam LM The type of the lower bound map. The default
164 171
     map type is \ref concepts::Digraph::ArcMap "GR::ArcMap<int>".
165 172
     \tparam UM The type of the upper bound (capacity) map.
166 173
     The default map type is \c LM.
167 174
     \tparam SM The type of the supply map. The default map type is
168 175
     \ref concepts::Digraph::NodeMap "GR::NodeMap<UM::Value>".
169 176
  */
170 177
#ifdef DOXYGEN
171 178
template< typename GR,
172 179
          typename LM,
173 180
          typename UM,
174 181
          typename SM,
175 182
          typename TR >
176 183
#else
177 184
template< typename GR,
178 185
          typename LM = typename GR::template ArcMap<int>,
179 186
          typename UM = LM,
180 187
          typename SM = typename GR::template NodeMap<typename UM::Value>,
181 188
          typename TR = CirculationDefaultTraits<GR, LM, UM, SM> >
182 189
#endif
183 190
  class Circulation {
184 191
  public:
185 192

	
186 193
    ///The \ref CirculationDefaultTraits "traits class" of the algorithm.
187 194
    typedef TR Traits;
188 195
    ///The type of the digraph the algorithm runs on.
... ...
@@ -357,211 +364,213 @@
357 364
        _local_flow = true;
358 365
      }
359 366
      if (!_level) {
360 367
        _level = Traits::createElevator(_g, _node_num);
361 368
        _local_level = true;
362 369
      }
363 370
      if (!_excess) {
364 371
        _excess = new ExcessMap(_g);
365 372
      }
366 373
    }
367 374

	
368 375
    void destroyStructures() {
369 376
      if (_local_flow) {
370 377
        delete _flow;
371 378
      }
372 379
      if (_local_level) {
373 380
        delete _level;
374 381
      }
375 382
      if (_excess) {
376 383
        delete _excess;
377 384
      }
378 385
    }
379 386

	
380 387
  public:
381 388

	
382 389
    /// Sets the lower bound map.
383 390

	
384 391
    /// Sets the lower bound map.
385 392
    /// \return <tt>(*this)</tt>
386 393
    Circulation& lowerMap(const LowerMap& map) {
387 394
      _lo = &map;
388 395
      return *this;
389 396
    }
390 397

	
391 398
    /// Sets the upper bound (capacity) map.
392 399

	
393 400
    /// Sets the upper bound (capacity) map.
394 401
    /// \return <tt>(*this)</tt>
395 402
    Circulation& upperMap(const UpperMap& map) {
396 403
      _up = &map;
397 404
      return *this;
398 405
    }
399 406

	
400 407
    /// Sets the supply map.
401 408

	
402 409
    /// Sets the supply map.
403 410
    /// \return <tt>(*this)</tt>
404 411
    Circulation& supplyMap(const SupplyMap& map) {
405 412
      _supply = &map;
406 413
      return *this;
407 414
    }
408 415

	
409 416
    /// \brief Sets the flow map.
410 417
    ///
411 418
    /// Sets the flow map.
412 419
    /// If you don't use this function before calling \ref run() or
413 420
    /// \ref init(), an instance will be allocated automatically.
414 421
    /// The destructor deallocates this automatically allocated map,
415 422
    /// of course.
416 423
    /// \return <tt>(*this)</tt>
417 424
    Circulation& flowMap(FlowMap& map) {
418 425
      if (_local_flow) {
419 426
        delete _flow;
420 427
        _local_flow = false;
421 428
      }
422 429
      _flow = &map;
423 430
      return *this;
424 431
    }
425 432

	
426 433
    /// \brief Sets the elevator used by algorithm.
427 434
    ///
428 435
    /// Sets the elevator used by algorithm.
429 436
    /// If you don't use this function before calling \ref run() or
430 437
    /// \ref init(), an instance will be allocated automatically.
431 438
    /// The destructor deallocates this automatically allocated elevator,
432 439
    /// of course.
433 440
    /// \return <tt>(*this)</tt>
434 441
    Circulation& elevator(Elevator& elevator) {
435 442
      if (_local_level) {
436 443
        delete _level;
437 444
        _local_level = false;
438 445
      }
439 446
      _level = &elevator;
440 447
      return *this;
441 448
    }
442 449

	
443 450
    /// \brief Returns a const reference to the elevator.
444 451
    ///
445 452
    /// Returns a const reference to the elevator.
446 453
    ///
447 454
    /// \pre Either \ref run() or \ref init() must be called before
448 455
    /// using this function.
449 456
    const Elevator& elevator() const {
450 457
      return *_level;
451 458
    }
452 459

	
453
    /// \brief Sets the tolerance used by algorithm.
460
    /// \brief Sets the tolerance used by the algorithm.
454 461
    ///
455
    /// Sets the tolerance used by algorithm.
456
    Circulation& tolerance(const Tolerance& tolerance) const {
462
    /// Sets the tolerance object used by the algorithm.
463
    /// \return <tt>(*this)</tt>
464
    Circulation& tolerance(const Tolerance& tolerance) {
457 465
      _tol = tolerance;
458 466
      return *this;
459 467
    }
460 468

	
461 469
    /// \brief Returns a const reference to the tolerance.
462 470
    ///
463
    /// Returns a const reference to the tolerance.
471
    /// Returns a const reference to the tolerance object used by
472
    /// the algorithm.
464 473
    const Tolerance& tolerance() const {
465
      return tolerance;
474
      return _tol;
466 475
    }
467 476

	
468 477
    /// \name Execution Control
469 478
    /// The simplest way to execute the algorithm is to call \ref run().\n
470
    /// If you need more control on the initial solution or the execution,
471
    /// first you have to call one of the \ref init() functions, then
479
    /// If you need better control on the initial solution or the execution,
480
    /// you have to call one of the \ref init() functions first, then
472 481
    /// the \ref start() function.
473 482

	
474 483
    ///@{
475 484

	
476 485
    /// Initializes the internal data structures.
477 486

	
478 487
    /// Initializes the internal data structures and sets all flow values
479 488
    /// to the lower bound.
480 489
    void init()
481 490
    {
482 491
      LEMON_DEBUG(checkBoundMaps(),
483 492
        "Upper bounds must be greater or equal to the lower bounds");
484 493

	
485 494
      createStructures();
486 495

	
487 496
      for(NodeIt n(_g);n!=INVALID;++n) {
488 497
        (*_excess)[n] = (*_supply)[n];
489 498
      }
490 499

	
491 500
      for (ArcIt e(_g);e!=INVALID;++e) {
492 501
        _flow->set(e, (*_lo)[e]);
493 502
        (*_excess)[_g.target(e)] += (*_flow)[e];
494 503
        (*_excess)[_g.source(e)] -= (*_flow)[e];
495 504
      }
496 505

	
497 506
      // global relabeling tested, but in general case it provides
498 507
      // worse performance for random digraphs
499 508
      _level->initStart();
500 509
      for(NodeIt n(_g);n!=INVALID;++n)
501 510
        _level->initAddItem(n);
502 511
      _level->initFinish();
503 512
      for(NodeIt n(_g);n!=INVALID;++n)
504 513
        if(_tol.positive((*_excess)[n]))
505 514
          _level->activate(n);
506 515
    }
507 516

	
508 517
    /// Initializes the internal data structures using a greedy approach.
509 518

	
510 519
    /// Initializes the internal data structures using a greedy approach
511 520
    /// to construct the initial solution.
512 521
    void greedyInit()
513 522
    {
514 523
      LEMON_DEBUG(checkBoundMaps(),
515 524
        "Upper bounds must be greater or equal to the lower bounds");
516 525

	
517 526
      createStructures();
518 527

	
519 528
      for(NodeIt n(_g);n!=INVALID;++n) {
520 529
        (*_excess)[n] = (*_supply)[n];
521 530
      }
522 531

	
523 532
      for (ArcIt e(_g);e!=INVALID;++e) {
524 533
        if (!_tol.less(-(*_excess)[_g.target(e)], (*_up)[e])) {
525 534
          _flow->set(e, (*_up)[e]);
526 535
          (*_excess)[_g.target(e)] += (*_up)[e];
527 536
          (*_excess)[_g.source(e)] -= (*_up)[e];
528 537
        } else if (_tol.less(-(*_excess)[_g.target(e)], (*_lo)[e])) {
529 538
          _flow->set(e, (*_lo)[e]);
530 539
          (*_excess)[_g.target(e)] += (*_lo)[e];
531 540
          (*_excess)[_g.source(e)] -= (*_lo)[e];
532 541
        } else {
533 542
          Value fc = -(*_excess)[_g.target(e)];
534 543
          _flow->set(e, fc);
535 544
          (*_excess)[_g.target(e)] = 0;
536 545
          (*_excess)[_g.source(e)] -= fc;
537 546
        }
538 547
      }
539 548

	
540 549
      _level->initStart();
541 550
      for(NodeIt n(_g);n!=INVALID;++n)
542 551
        _level->initAddItem(n);
543 552
      _level->initFinish();
544 553
      for(NodeIt n(_g);n!=INVALID;++n)
545 554
        if(_tol.positive((*_excess)[n]))
546 555
          _level->activate(n);
547 556
    }
548 557

	
549 558
    ///Executes the algorithm
550 559

	
551 560
    ///This function executes the algorithm.
552 561
    ///
553 562
    ///\return \c true if a feasible circulation is found.
554 563
    ///
555 564
    ///\sa barrier()
556 565
    ///\sa barrierMap()
557 566
    bool start()
558 567
    {
559 568

	
560 569
      Node act;
561 570
      Node bact=INVALID;
562 571
      Node last_activated=INVALID;
563 572
      while((act=_level->highestActive())!=INVALID) {
564 573
        int actlevel=(*_level)[act];
565 574
        int mlevel=_node_num;
566 575
        Value exc=(*_excess)[act];
567 576

	
Ignore white space 6 line context
1 1
/* -*- mode: C++; indent-tabs-mode: nil; -*-
2 2
 *
3 3
 * This file is a part of LEMON, a generic C++ optimization library.
4 4
 *
5 5
 * Copyright (C) 2003-2009
6 6
 * Egervary Jeno Kombinatorikus Optimalizalasi Kutatocsoport
7 7
 * (Egervary Research Group on Combinatorial Optimization, EGRES).
8 8
 *
9 9
 * Permission to use, modify and distribute this software is granted
10 10
 * provided that this copyright notice appears in all copies. For
11 11
 * precise terms see the accompanying LICENSE file.
12 12
 *
13 13
 * This software is provided "AS IS" with no warranty of any kind,
14 14
 * express or implied, and with no claim as to its suitability for any
15 15
 * purpose.
16 16
 *
17 17
 */
18 18

	
19
#ifndef LEMON_CONCEPTS_HEAP_H
20
#define LEMON_CONCEPTS_HEAP_H
21

	
19 22
///\ingroup concept
20 23
///\file
21 24
///\brief The concept of heaps.
22 25

	
23
#ifndef LEMON_CONCEPTS_HEAP_H
24
#define LEMON_CONCEPTS_HEAP_H
25

	
26 26
#include <lemon/core.h>
27 27
#include <lemon/concept_check.h>
28 28

	
29 29
namespace lemon {
30 30

	
31 31
  namespace concepts {
32 32

	
33 33
    /// \addtogroup concept
34 34
    /// @{
35 35

	
36 36
    /// \brief The heap concept.
37 37
    ///
38
    /// Concept class describing the main interface of heaps. A \e heap
39
    /// is a data structure for storing items with specified values called
40
    /// \e priorities in such a way that finding the item with minimum
41
    /// priority is efficient. In a heap one can change the priority of an
42
    /// item, add or erase an item, etc.
38
    /// This concept class describes the main interface of heaps.
39
    /// The various \ref heaps "heap structures" are efficient
40
    /// implementations of the abstract data type \e priority \e queue.
41
    /// They store items with specified values called \e priorities
42
    /// in such a way that finding and removing the item with minimum
43
    /// priority are efficient. The basic operations are adding and
44
    /// erasing items, changing the priority of an item, etc.
43 45
    ///
44
    /// \tparam PR Type of the priority of the items.
45
    /// \tparam IM A read and writable item map with int values, used
46
    /// Heaps are crucial in several algorithms, such as Dijkstra and Prim.
47
    /// Any class that conforms to this concept can be used easily in such
48
    /// algorithms.
49
    ///
50
    /// \tparam PR Type of the priorities of the items.
51
    /// \tparam IM A read-writable item map with \c int values, used
46 52
    /// internally to handle the cross references.
47
    /// \tparam Comp A functor class for the ordering of the priorities.
53
    /// \tparam CMP A functor class for comparing the priorities.
48 54
    /// The default is \c std::less<PR>.
49 55
#ifdef DOXYGEN
50
    template <typename PR, typename IM, typename Comp = std::less<PR> >
56
    template <typename PR, typename IM, typename CMP>
51 57
#else
52
    template <typename PR, typename IM>
58
    template <typename PR, typename IM, typename CMP = std::less<PR> >
53 59
#endif
54 60
    class Heap {
55 61
    public:
56 62

	
57 63
      /// Type of the item-int map.
58 64
      typedef IM ItemIntMap;
59 65
      /// Type of the priorities.
60 66
      typedef PR Prio;
61 67
      /// Type of the items stored in the heap.
62 68
      typedef typename ItemIntMap::Key Item;
63 69

	
64 70
      /// \brief Type to represent the states of the items.
65 71
      ///
66 72
      /// Each item has a state associated to it. It can be "in heap",
67
      /// "pre heap" or "post heap". The later two are indifferent
68
      /// from the point of view of the heap, but may be useful for
69
      /// the user.
73
      /// "pre-heap" or "post-heap". The latter two are indifferent from the
74
      /// heap's point of view, but may be useful to the user.
70 75
      ///
71 76
      /// The item-int map must be initialized in such way that it assigns
72 77
      /// \c PRE_HEAP (<tt>-1</tt>) to any element to be put in the heap.
73 78
      enum State {
74 79
        IN_HEAP = 0,    ///< = 0. The "in heap" state constant.
75
        PRE_HEAP = -1,  ///< = -1. The "pre heap" state constant.
76
        POST_HEAP = -2  ///< = -2. The "post heap" state constant.
80
        PRE_HEAP = -1,  ///< = -1. The "pre-heap" state constant.
81
        POST_HEAP = -2  ///< = -2. The "post-heap" state constant.
77 82
      };
78 83

	
79
      /// \brief The constructor.
84
      /// \brief Constructor.
80 85
      ///
81
      /// The constructor.
86
      /// Constructor.
82 87
      /// \param map A map that assigns \c int values to keys of type
83 88
      /// \c Item. It is used internally by the heap implementations to
84 89
      /// handle the cross references. The assigned value must be
85
      /// \c PRE_HEAP (<tt>-1</tt>) for every item.
90
      /// \c PRE_HEAP (<tt>-1</tt>) for each item.
86 91
      explicit Heap(ItemIntMap &map) {}
87 92

	
93
      /// \brief Constructor.
94
      ///
95
      /// Constructor.
96
      /// \param map A map that assigns \c int values to keys of type
97
      /// \c Item. It is used internally by the heap implementations to
98
      /// handle the cross references. The assigned value must be
99
      /// \c PRE_HEAP (<tt>-1</tt>) for each item.
100
      /// \param comp The function object used for comparing the priorities.
101
      explicit Heap(ItemIntMap &map, const CMP &comp) {}
102

	
88 103
      /// \brief The number of items stored in the heap.
89 104
      ///
90
      /// Returns the number of items stored in the heap.
105
      /// This function returns the number of items stored in the heap.
91 106
      int size() const { return 0; }
92 107

	
93
      /// \brief Checks if the heap is empty.
108
      /// \brief Check if the heap is empty.
94 109
      ///
95
      /// Returns \c true if the heap is empty.
110
      /// This function returns \c true if the heap is empty.
96 111
      bool empty() const { return false; }
97 112

	
98
      /// \brief Makes the heap empty.
113
      /// \brief Make the heap empty.
99 114
      ///
100
      /// Makes the heap empty.
101
      void clear();
115
      /// This functon makes the heap empty.
116
      /// It does not change the cross reference map. If you want to reuse
117
      /// a heap that is not surely empty, you should first clear it and
118
      /// then you should set the cross reference map to \c PRE_HEAP
119
      /// for each item.
120
      void clear() {}
102 121

	
103
      /// \brief Inserts an item into the heap with the given priority.
122
      /// \brief Insert an item into the heap with the given priority.
104 123
      ///
105
      /// Inserts the given item into the heap with the given priority.
124
      /// This function inserts the given item into the heap with the
125
      /// given priority.
106 126
      /// \param i The item to insert.
107 127
      /// \param p The priority of the item.
128
      /// \pre \e i must not be stored in the heap.
108 129
      void push(const Item &i, const Prio &p) {}
109 130

	
110
      /// \brief Returns the item having minimum priority.
131
      /// \brief Return the item having minimum priority.
111 132
      ///
112
      /// Returns the item having minimum priority.
133
      /// This function returns the item having minimum priority.
113 134
      /// \pre The heap must be non-empty.
114 135
      Item top() const {}
115 136

	
116 137
      /// \brief The minimum priority.
117 138
      ///
118
      /// Returns the minimum priority.
139
      /// This function returns the minimum priority.
119 140
      /// \pre The heap must be non-empty.
120 141
      Prio prio() const {}
121 142

	
122
      /// \brief Removes the item having minimum priority.
143
      /// \brief Remove the item having minimum priority.
123 144
      ///
124
      /// Removes the item having minimum priority.
145
      /// This function removes the item having minimum priority.
125 146
      /// \pre The heap must be non-empty.
126 147
      void pop() {}
127 148

	
128
      /// \brief Removes an item from the heap.
149
      /// \brief Remove the given item from the heap.
129 150
      ///
130
      /// Removes the given item from the heap if it is already stored.
151
      /// This function removes the given item from the heap if it is
152
      /// already stored.
131 153
      /// \param i The item to delete.
154
      /// \pre \e i must be in the heap.
132 155
      void erase(const Item &i) {}
133 156

	
134
      /// \brief The priority of an item.
157
      /// \brief The priority of the given item.
135 158
      ///
136
      /// Returns the priority of the given item.
159
      /// This function returns the priority of the given item.
137 160
      /// \param i The item.
138
      /// \pre \c i must be in the heap.
161
      /// \pre \e i must be in the heap.
139 162
      Prio operator[](const Item &i) const {}
140 163

	
141
      /// \brief Sets the priority of an item or inserts it, if it is
164
      /// \brief Set the priority of an item or insert it, if it is
142 165
      /// not stored in the heap.
143 166
      ///
144 167
      /// This method sets the priority of the given item if it is
145
      /// already stored in the heap.
146
      /// Otherwise it inserts the given item with the given priority.
168
      /// already stored in the heap. Otherwise it inserts the given
169
      /// item into the heap with the given priority.
147 170
      ///
148 171
      /// \param i The item.
149 172
      /// \param p The priority.
150 173
      void set(const Item &i, const Prio &p) {}
151 174

	
152
      /// \brief Decreases the priority of an item to the given value.
175
      /// \brief Decrease the priority of an item to the given value.
153 176
      ///
154
      /// Decreases the priority of an item to the given value.
177
      /// This function decreases the priority of an item to the given value.
155 178
      /// \param i The item.
156 179
      /// \param p The priority.
157
      /// \pre \c i must be stored in the heap with priority at least \c p.
180
      /// \pre \e i must be stored in the heap with priority at least \e p.
158 181
      void decrease(const Item &i, const Prio &p) {}
159 182

	
160
      /// \brief Increases the priority of an item to the given value.
183
      /// \brief Increase the priority of an item to the given value.
161 184
      ///
162
      /// Increases the priority of an item to the given value.
185
      /// This function increases the priority of an item to the given value.
163 186
      /// \param i The item.
164 187
      /// \param p The priority.
165
      /// \pre \c i must be stored in the heap with priority at most \c p.
188
      /// \pre \e i must be stored in the heap with priority at most \e p.
166 189
      void increase(const Item &i, const Prio &p) {}
167 190

	
168
      /// \brief Returns if an item is in, has already been in, or has
169
      /// never been in the heap.
191
      /// \brief Return the state of an item.
170 192
      ///
171 193
      /// This method returns \c PRE_HEAP if the given item has never
172 194
      /// been in the heap, \c IN_HEAP if it is in the heap at the moment,
173 195
      /// and \c POST_HEAP otherwise.
174 196
      /// In the latter case it is possible that the item will get back
175 197
      /// to the heap again.
176 198
      /// \param i The item.
177 199
      State state(const Item &i) const {}
178 200

	
179
      /// \brief Sets the state of an item in the heap.
201
      /// \brief Set the state of an item in the heap.
180 202
      ///
181
      /// Sets the state of the given item in the heap. It can be used
182
      /// to manually clear the heap when it is important to achive the
183
      /// better time complexity.
203
      /// This function sets the state of the given item in the heap.
204
      /// It can be used to manually clear the heap when it is important
205
      /// to achive better time complexity.
184 206
      /// \param i The item.
185 207
      /// \param st The state. It should not be \c IN_HEAP.
186 208
      void state(const Item& i, State st) {}
187 209

	
188 210

	
189 211
      template <typename _Heap>
190 212
      struct Constraints {
191 213
      public:
192 214
        void constraints() {
193 215
          typedef typename _Heap::Item OwnItem;
194 216
          typedef typename _Heap::Prio OwnPrio;
195 217
          typedef typename _Heap::State OwnState;
196 218

	
197 219
          Item item;
198 220
          Prio prio;
199 221
          item=Item();
200 222
          prio=Prio();
201 223
          ignore_unused_variable_warning(item);
202 224
          ignore_unused_variable_warning(prio);
203 225

	
204 226
          OwnItem own_item;
205 227
          OwnPrio own_prio;
206 228
          OwnState own_state;
207 229
          own_item=Item();
208 230
          own_prio=Prio();
209 231
          ignore_unused_variable_warning(own_item);
210 232
          ignore_unused_variable_warning(own_prio);
211 233
          ignore_unused_variable_warning(own_state);
212 234

	
213 235
          _Heap heap1(map);
214 236
          _Heap heap2 = heap1;
215 237
          ignore_unused_variable_warning(heap1);
216 238
          ignore_unused_variable_warning(heap2);
217 239

	
218 240
          int s = heap.size();
219 241
          ignore_unused_variable_warning(s);
220 242
          bool e = heap.empty();
221 243
          ignore_unused_variable_warning(e);
222 244

	
223 245
          prio = heap.prio();
224 246
          item = heap.top();
225 247
          prio = heap[item];
226 248
          own_prio = heap.prio();
227 249
          own_item = heap.top();
228 250
          own_prio = heap[own_item];
229 251

	
230 252
          heap.push(item, prio);
231 253
          heap.push(own_item, own_prio);
232 254
          heap.pop();
233 255

	
234 256
          heap.set(item, prio);
235 257
          heap.decrease(item, prio);
236 258
          heap.increase(item, prio);
237 259
          heap.set(own_item, own_prio);
238 260
          heap.decrease(own_item, own_prio);
239 261
          heap.increase(own_item, own_prio);
240 262

	
241 263
          heap.erase(item);
242 264
          heap.erase(own_item);
243 265
          heap.clear();
244 266

	
245 267
          own_state = heap.state(own_item);
246 268
          heap.state(own_item, own_state);
247 269

	
248 270
          own_state = _Heap::PRE_HEAP;
249 271
          own_state = _Heap::IN_HEAP;
250 272
          own_state = _Heap::POST_HEAP;
251 273
        }
252 274

	
253 275
        _Heap& heap;
254 276
        ItemIntMap& map;
255 277
      };
256 278
    };
257 279

	
258 280
    /// @}
259 281
  } // namespace lemon
260 282
}
261 283
#endif
Ignore white space 6 line context
... ...
@@ -319,194 +319,194 @@
319 319
    ///\param g The digraph the algorithm runs on.
320 320
    Dfs(const Digraph &g) :
321 321
      G(&g),
322 322
      _pred(NULL), local_pred(false),
323 323
      _dist(NULL), local_dist(false),
324 324
      _reached(NULL), local_reached(false),
325 325
      _processed(NULL), local_processed(false)
326 326
    { }
327 327

	
328 328
    ///Destructor.
329 329
    ~Dfs()
330 330
    {
331 331
      if(local_pred) delete _pred;
332 332
      if(local_dist) delete _dist;
333 333
      if(local_reached) delete _reached;
334 334
      if(local_processed) delete _processed;
335 335
    }
336 336

	
337 337
    ///Sets the map that stores the predecessor arcs.
338 338

	
339 339
    ///Sets the map that stores the predecessor arcs.
340 340
    ///If you don't use this function before calling \ref run(Node) "run()"
341 341
    ///or \ref init(), an instance will be allocated automatically.
342 342
    ///The destructor deallocates this automatically allocated map,
343 343
    ///of course.
344 344
    ///\return <tt> (*this) </tt>
345 345
    Dfs &predMap(PredMap &m)
346 346
    {
347 347
      if(local_pred) {
348 348
        delete _pred;
349 349
        local_pred=false;
350 350
      }
351 351
      _pred = &m;
352 352
      return *this;
353 353
    }
354 354

	
355 355
    ///Sets the map that indicates which nodes are reached.
356 356

	
357 357
    ///Sets the map that indicates which nodes are reached.
358 358
    ///If you don't use this function before calling \ref run(Node) "run()"
359 359
    ///or \ref init(), an instance will be allocated automatically.
360 360
    ///The destructor deallocates this automatically allocated map,
361 361
    ///of course.
362 362
    ///\return <tt> (*this) </tt>
363 363
    Dfs &reachedMap(ReachedMap &m)
364 364
    {
365 365
      if(local_reached) {
366 366
        delete _reached;
367 367
        local_reached=false;
368 368
      }
369 369
      _reached = &m;
370 370
      return *this;
371 371
    }
372 372

	
373 373
    ///Sets the map that indicates which nodes are processed.
374 374

	
375 375
    ///Sets the map that indicates which nodes are processed.
376 376
    ///If you don't use this function before calling \ref run(Node) "run()"
377 377
    ///or \ref init(), an instance will be allocated automatically.
378 378
    ///The destructor deallocates this automatically allocated map,
379 379
    ///of course.
380 380
    ///\return <tt> (*this) </tt>
381 381
    Dfs &processedMap(ProcessedMap &m)
382 382
    {
383 383
      if(local_processed) {
384 384
        delete _processed;
385 385
        local_processed=false;
386 386
      }
387 387
      _processed = &m;
388 388
      return *this;
389 389
    }
390 390

	
391 391
    ///Sets the map that stores the distances of the nodes.
392 392

	
393 393
    ///Sets the map that stores the distances of the nodes calculated by
394 394
    ///the algorithm.
395 395
    ///If you don't use this function before calling \ref run(Node) "run()"
396 396
    ///or \ref init(), an instance will be allocated automatically.
397 397
    ///The destructor deallocates this automatically allocated map,
398 398
    ///of course.
399 399
    ///\return <tt> (*this) </tt>
400 400
    Dfs &distMap(DistMap &m)
401 401
    {
402 402
      if(local_dist) {
403 403
        delete _dist;
404 404
        local_dist=false;
405 405
      }
406 406
      _dist = &m;
407 407
      return *this;
408 408
    }
409 409

	
410 410
  public:
411 411

	
412 412
    ///\name Execution Control
413 413
    ///The simplest way to execute the DFS algorithm is to use one of the
414 414
    ///member functions called \ref run(Node) "run()".\n
415
    ///If you need more control on the execution, first you have to call
416
    ///\ref init(), then you can add a source node with \ref addSource()
415
    ///If you need better control on the execution, you have to call
416
    ///\ref init() first, then you can add a source node with \ref addSource()
417 417
    ///and perform the actual computation with \ref start().
418 418
    ///This procedure can be repeated if there are nodes that have not
419 419
    ///been reached.
420 420

	
421 421
    ///@{
422 422

	
423 423
    ///\brief Initializes the internal data structures.
424 424
    ///
425 425
    ///Initializes the internal data structures.
426 426
    void init()
427 427
    {
428 428
      create_maps();
429 429
      _stack.resize(countNodes(*G));
430 430
      _stack_head=-1;
431 431
      for ( NodeIt u(*G) ; u!=INVALID ; ++u ) {
432 432
        _pred->set(u,INVALID);
433 433
        _reached->set(u,false);
434 434
        _processed->set(u,false);
435 435
      }
436 436
    }
437 437

	
438 438
    ///Adds a new source node.
439 439

	
440 440
    ///Adds a new source node to the set of nodes to be processed.
441 441
    ///
442 442
    ///\pre The stack must be empty. Otherwise the algorithm gives
443 443
    ///wrong results. (One of the outgoing arcs of all the source nodes
444 444
    ///except for the last one will not be visited and distances will
445 445
    ///also be wrong.)
446 446
    void addSource(Node s)
447 447
    {
448 448
      LEMON_DEBUG(emptyQueue(), "The stack is not empty.");
449 449
      if(!(*_reached)[s])
450 450
        {
451 451
          _reached->set(s,true);
452 452
          _pred->set(s,INVALID);
453 453
          OutArcIt e(*G,s);
454 454
          if(e!=INVALID) {
455 455
            _stack[++_stack_head]=e;
456 456
            _dist->set(s,_stack_head);
457 457
          }
458 458
          else {
459 459
            _processed->set(s,true);
460 460
            _dist->set(s,0);
461 461
          }
462 462
        }
463 463
    }
464 464

	
465 465
    ///Processes the next arc.
466 466

	
467 467
    ///Processes the next arc.
468 468
    ///
469 469
    ///\return The processed arc.
470 470
    ///
471 471
    ///\pre The stack must not be empty.
472 472
    Arc processNextArc()
473 473
    {
474 474
      Node m;
475 475
      Arc e=_stack[_stack_head];
476 476
      if(!(*_reached)[m=G->target(e)]) {
477 477
        _pred->set(m,e);
478 478
        _reached->set(m,true);
479 479
        ++_stack_head;
480 480
        _stack[_stack_head] = OutArcIt(*G, m);
481 481
        _dist->set(m,_stack_head);
482 482
      }
483 483
      else {
484 484
        m=G->source(e);
485 485
        ++_stack[_stack_head];
486 486
      }
487 487
      while(_stack_head>=0 && _stack[_stack_head]==INVALID) {
488 488
        _processed->set(m,true);
489 489
        --_stack_head;
490 490
        if(_stack_head>=0) {
491 491
          m=G->source(_stack[_stack_head]);
492 492
          ++_stack[_stack_head];
493 493
        }
494 494
      }
495 495
      return e;
496 496
    }
497 497

	
498 498
    ///Next arc to be processed.
499 499

	
500 500
    ///Next arc to be processed.
501 501
    ///
502 502
    ///\return The next arc to be processed or \c INVALID if the stack
503 503
    ///is empty.
504 504
    OutArcIt nextArc() const
505 505
    {
506 506
      return _stack_head>=0?_stack[_stack_head]:INVALID;
507 507
    }
508 508

	
509 509
    ///Returns \c false if there are nodes to be processed.
510 510

	
511 511
    ///Returns \c false if there are nodes to be processed
512 512
    ///in the queue (stack).
... ...
@@ -1271,194 +1271,194 @@
1271 1271
  private:
1272 1272

	
1273 1273
    typedef typename Digraph::Node Node;
1274 1274
    typedef typename Digraph::NodeIt NodeIt;
1275 1275
    typedef typename Digraph::Arc Arc;
1276 1276
    typedef typename Digraph::OutArcIt OutArcIt;
1277 1277

	
1278 1278
    //Pointer to the underlying digraph.
1279 1279
    const Digraph *_digraph;
1280 1280
    //Pointer to the visitor object.
1281 1281
    Visitor *_visitor;
1282 1282
    //Pointer to the map of reached status of the nodes.
1283 1283
    ReachedMap *_reached;
1284 1284
    //Indicates if _reached is locally allocated (true) or not.
1285 1285
    bool local_reached;
1286 1286

	
1287 1287
    std::vector<typename Digraph::Arc> _stack;
1288 1288
    int _stack_head;
1289 1289

	
1290 1290
    //Creates the maps if necessary.
1291 1291
    void create_maps() {
1292 1292
      if(!_reached) {
1293 1293
        local_reached = true;
1294 1294
        _reached = Traits::createReachedMap(*_digraph);
1295 1295
      }
1296 1296
    }
1297 1297

	
1298 1298
  protected:
1299 1299

	
1300 1300
    DfsVisit() {}
1301 1301

	
1302 1302
  public:
1303 1303

	
1304 1304
    typedef DfsVisit Create;
1305 1305

	
1306 1306
    /// \name Named Template Parameters
1307 1307

	
1308 1308
    ///@{
1309 1309
    template <class T>
1310 1310
    struct SetReachedMapTraits : public Traits {
1311 1311
      typedef T ReachedMap;
1312 1312
      static ReachedMap *createReachedMap(const Digraph &digraph) {
1313 1313
        LEMON_ASSERT(false, "ReachedMap is not initialized");
1314 1314
        return 0; // ignore warnings
1315 1315
      }
1316 1316
    };
1317 1317
    /// \brief \ref named-templ-param "Named parameter" for setting
1318 1318
    /// ReachedMap type.
1319 1319
    ///
1320 1320
    /// \ref named-templ-param "Named parameter" for setting ReachedMap type.
1321 1321
    template <class T>
1322 1322
    struct SetReachedMap : public DfsVisit< Digraph, Visitor,
1323 1323
                                            SetReachedMapTraits<T> > {
1324 1324
      typedef DfsVisit< Digraph, Visitor, SetReachedMapTraits<T> > Create;
1325 1325
    };
1326 1326
    ///@}
1327 1327

	
1328 1328
  public:
1329 1329

	
1330 1330
    /// \brief Constructor.
1331 1331
    ///
1332 1332
    /// Constructor.
1333 1333
    ///
1334 1334
    /// \param digraph The digraph the algorithm runs on.
1335 1335
    /// \param visitor The visitor object of the algorithm.
1336 1336
    DfsVisit(const Digraph& digraph, Visitor& visitor)
1337 1337
      : _digraph(&digraph), _visitor(&visitor),
1338 1338
        _reached(0), local_reached(false) {}
1339 1339

	
1340 1340
    /// \brief Destructor.
1341 1341
    ~DfsVisit() {
1342 1342
      if(local_reached) delete _reached;
1343 1343
    }
1344 1344

	
1345 1345
    /// \brief Sets the map that indicates which nodes are reached.
1346 1346
    ///
1347 1347
    /// Sets the map that indicates which nodes are reached.
1348 1348
    /// If you don't use this function before calling \ref run(Node) "run()"
1349 1349
    /// or \ref init(), an instance will be allocated automatically.
1350 1350
    /// The destructor deallocates this automatically allocated map,
1351 1351
    /// of course.
1352 1352
    /// \return <tt> (*this) </tt>
1353 1353
    DfsVisit &reachedMap(ReachedMap &m) {
1354 1354
      if(local_reached) {
1355 1355
        delete _reached;
1356 1356
        local_reached=false;
1357 1357
      }
1358 1358
      _reached = &m;
1359 1359
      return *this;
1360 1360
    }
1361 1361

	
1362 1362
  public:
1363 1363

	
1364 1364
    /// \name Execution Control
1365 1365
    /// The simplest way to execute the DFS algorithm is to use one of the
1366 1366
    /// member functions called \ref run(Node) "run()".\n
1367
    /// If you need more control on the execution, first you have to call
1368
    /// \ref init(), then you can add a source node with \ref addSource()
1367
    /// If you need better control on the execution, you have to call
1368
    /// \ref init() first, then you can add a source node with \ref addSource()
1369 1369
    /// and perform the actual computation with \ref start().
1370 1370
    /// This procedure can be repeated if there are nodes that have not
1371 1371
    /// been reached.
1372 1372

	
1373 1373
    /// @{
1374 1374

	
1375 1375
    /// \brief Initializes the internal data structures.
1376 1376
    ///
1377 1377
    /// Initializes the internal data structures.
1378 1378
    void init() {
1379 1379
      create_maps();
1380 1380
      _stack.resize(countNodes(*_digraph));
1381 1381
      _stack_head = -1;
1382 1382
      for (NodeIt u(*_digraph) ; u != INVALID ; ++u) {
1383 1383
        _reached->set(u, false);
1384 1384
      }
1385 1385
    }
1386 1386

	
1387 1387
    /// \brief Adds a new source node.
1388 1388
    ///
1389 1389
    /// Adds a new source node to the set of nodes to be processed.
1390 1390
    ///
1391 1391
    /// \pre The stack must be empty. Otherwise the algorithm gives
1392 1392
    /// wrong results. (One of the outgoing arcs of all the source nodes
1393 1393
    /// except for the last one will not be visited and distances will
1394 1394
    /// also be wrong.)
1395 1395
    void addSource(Node s)
1396 1396
    {
1397 1397
      LEMON_DEBUG(emptyQueue(), "The stack is not empty.");
1398 1398
      if(!(*_reached)[s]) {
1399 1399
          _reached->set(s,true);
1400 1400
          _visitor->start(s);
1401 1401
          _visitor->reach(s);
1402 1402
          Arc e;
1403 1403
          _digraph->firstOut(e, s);
1404 1404
          if (e != INVALID) {
1405 1405
            _stack[++_stack_head] = e;
1406 1406
          } else {
1407 1407
            _visitor->leave(s);
1408 1408
            _visitor->stop(s);
1409 1409
          }
1410 1410
        }
1411 1411
    }
1412 1412

	
1413 1413
    /// \brief Processes the next arc.
1414 1414
    ///
1415 1415
    /// Processes the next arc.
1416 1416
    ///
1417 1417
    /// \return The processed arc.
1418 1418
    ///
1419 1419
    /// \pre The stack must not be empty.
1420 1420
    Arc processNextArc() {
1421 1421
      Arc e = _stack[_stack_head];
1422 1422
      Node m = _digraph->target(e);
1423 1423
      if(!(*_reached)[m]) {
1424 1424
        _visitor->discover(e);
1425 1425
        _visitor->reach(m);
1426 1426
        _reached->set(m, true);
1427 1427
        _digraph->firstOut(_stack[++_stack_head], m);
1428 1428
      } else {
1429 1429
        _visitor->examine(e);
1430 1430
        m = _digraph->source(e);
1431 1431
        _digraph->nextOut(_stack[_stack_head]);
1432 1432
      }
1433 1433
      while (_stack_head>=0 && _stack[_stack_head] == INVALID) {
1434 1434
        _visitor->leave(m);
1435 1435
        --_stack_head;
1436 1436
        if (_stack_head >= 0) {
1437 1437
          _visitor->backtrack(_stack[_stack_head]);
1438 1438
          m = _digraph->source(_stack[_stack_head]);
1439 1439
          _digraph->nextOut(_stack[_stack_head]);
1440 1440
        } else {
1441 1441
          _visitor->stop(m);
1442 1442
        }
1443 1443
      }
1444 1444
      return e;
1445 1445
    }
1446 1446

	
1447 1447
    /// \brief Next arc to be processed.
1448 1448
    ///
1449 1449
    /// Next arc to be processed.
1450 1450
    ///
1451 1451
    /// \return The next arc to be processed or INVALID if the stack is
1452 1452
    /// empty.
1453 1453
    Arc nextArc() const {
1454 1454
      return _stack_head >= 0 ? _stack[_stack_head] : INVALID;
1455 1455
    }
1456 1456

	
1457 1457
    /// \brief Returns \c false if there are nodes
1458 1458
    /// to be processed.
1459 1459
    ///
1460 1460
    /// Returns \c false if there are nodes
1461 1461
    /// to be processed in the queue (stack).
1462 1462
    bool emptyQueue() const { return _stack_head < 0; }
1463 1463

	
1464 1464
    /// \brief Returns the number of the nodes to be processed.
Ignore white space 6 line context
... ...
@@ -496,194 +496,194 @@
496 496
      _length = &m;
497 497
      return *this;
498 498
    }
499 499

	
500 500
    ///Sets the map that stores the predecessor arcs.
501 501

	
502 502
    ///Sets the map that stores the predecessor arcs.
503 503
    ///If you don't use this function before calling \ref run(Node) "run()"
504 504
    ///or \ref init(), an instance will be allocated automatically.
505 505
    ///The destructor deallocates this automatically allocated map,
506 506
    ///of course.
507 507
    ///\return <tt> (*this) </tt>
508 508
    Dijkstra &predMap(PredMap &m)
509 509
    {
510 510
      if(local_pred) {
511 511
        delete _pred;
512 512
        local_pred=false;
513 513
      }
514 514
      _pred = &m;
515 515
      return *this;
516 516
    }
517 517

	
518 518
    ///Sets the map that indicates which nodes are processed.
519 519

	
520 520
    ///Sets the map that indicates which nodes are processed.
521 521
    ///If you don't use this function before calling \ref run(Node) "run()"
522 522
    ///or \ref init(), an instance will be allocated automatically.
523 523
    ///The destructor deallocates this automatically allocated map,
524 524
    ///of course.
525 525
    ///\return <tt> (*this) </tt>
526 526
    Dijkstra &processedMap(ProcessedMap &m)
527 527
    {
528 528
      if(local_processed) {
529 529
        delete _processed;
530 530
        local_processed=false;
531 531
      }
532 532
      _processed = &m;
533 533
      return *this;
534 534
    }
535 535

	
536 536
    ///Sets the map that stores the distances of the nodes.
537 537

	
538 538
    ///Sets the map that stores the distances of the nodes calculated by the
539 539
    ///algorithm.
540 540
    ///If you don't use this function before calling \ref run(Node) "run()"
541 541
    ///or \ref init(), an instance will be allocated automatically.
542 542
    ///The destructor deallocates this automatically allocated map,
543 543
    ///of course.
544 544
    ///\return <tt> (*this) </tt>
545 545
    Dijkstra &distMap(DistMap &m)
546 546
    {
547 547
      if(local_dist) {
548 548
        delete _dist;
549 549
        local_dist=false;
550 550
      }
551 551
      _dist = &m;
552 552
      return *this;
553 553
    }
554 554

	
555 555
    ///Sets the heap and the cross reference used by algorithm.
556 556

	
557 557
    ///Sets the heap and the cross reference used by algorithm.
558 558
    ///If you don't use this function before calling \ref run(Node) "run()"
559 559
    ///or \ref init(), heap and cross reference instances will be
560 560
    ///allocated automatically.
561 561
    ///The destructor deallocates these automatically allocated objects,
562 562
    ///of course.
563 563
    ///\return <tt> (*this) </tt>
564 564
    Dijkstra &heap(Heap& hp, HeapCrossRef &cr)
565 565
    {
566 566
      if(local_heap_cross_ref) {
567 567
        delete _heap_cross_ref;
568 568
        local_heap_cross_ref=false;
569 569
      }
570 570
      _heap_cross_ref = &cr;
571 571
      if(local_heap) {
572 572
        delete _heap;
573 573
        local_heap=false;
574 574
      }
575 575
      _heap = &hp;
576 576
      return *this;
577 577
    }
578 578

	
579 579
  private:
580 580

	
581 581
    void finalizeNodeData(Node v,Value dst)
582 582
    {
583 583
      _processed->set(v,true);
584 584
      _dist->set(v, dst);
585 585
    }
586 586

	
587 587
  public:
588 588

	
589 589
    ///\name Execution Control
590 590
    ///The simplest way to execute the %Dijkstra algorithm is to use
591 591
    ///one of the member functions called \ref run(Node) "run()".\n
592
    ///If you need more control on the execution, first you have to call
593
    ///\ref init(), then you can add several source nodes with
592
    ///If you need better control on the execution, you have to call
593
    ///\ref init() first, then you can add several source nodes with
594 594
    ///\ref addSource(). Finally the actual path computation can be
595 595
    ///performed with one of the \ref start() functions.
596 596

	
597 597
    ///@{
598 598

	
599 599
    ///\brief Initializes the internal data structures.
600 600
    ///
601 601
    ///Initializes the internal data structures.
602 602
    void init()
603 603
    {
604 604
      create_maps();
605 605
      _heap->clear();
606 606
      for ( NodeIt u(*G) ; u!=INVALID ; ++u ) {
607 607
        _pred->set(u,INVALID);
608 608
        _processed->set(u,false);
609 609
        _heap_cross_ref->set(u,Heap::PRE_HEAP);
610 610
      }
611 611
    }
612 612

	
613 613
    ///Adds a new source node.
614 614

	
615 615
    ///Adds a new source node to the priority heap.
616 616
    ///The optional second parameter is the initial distance of the node.
617 617
    ///
618 618
    ///The function checks if the node has already been added to the heap and
619 619
    ///it is pushed to the heap only if either it was not in the heap
620 620
    ///or the shortest path found till then is shorter than \c dst.
621 621
    void addSource(Node s,Value dst=OperationTraits::zero())
622 622
    {
623 623
      if(_heap->state(s) != Heap::IN_HEAP) {
624 624
        _heap->push(s,dst);
625 625
      } else if(OperationTraits::less((*_heap)[s], dst)) {
626 626
        _heap->set(s,dst);
627 627
        _pred->set(s,INVALID);
628 628
      }
629 629
    }
630 630

	
631 631
    ///Processes the next node in the priority heap
632 632

	
633 633
    ///Processes the next node in the priority heap.
634 634
    ///
635 635
    ///\return The processed node.
636 636
    ///
637 637
    ///\warning The priority heap must not be empty.
638 638
    Node processNextNode()
639 639
    {
640 640
      Node v=_heap->top();
641 641
      Value oldvalue=_heap->prio();
642 642
      _heap->pop();
643 643
      finalizeNodeData(v,oldvalue);
644 644

	
645 645
      for(OutArcIt e(*G,v); e!=INVALID; ++e) {
646 646
        Node w=G->target(e);
647 647
        switch(_heap->state(w)) {
648 648
        case Heap::PRE_HEAP:
649 649
          _heap->push(w,OperationTraits::plus(oldvalue, (*_length)[e]));
650 650
          _pred->set(w,e);
651 651
          break;
652 652
        case Heap::IN_HEAP:
653 653
          {
654 654
            Value newvalue = OperationTraits::plus(oldvalue, (*_length)[e]);
655 655
            if ( OperationTraits::less(newvalue, (*_heap)[w]) ) {
656 656
              _heap->decrease(w, newvalue);
657 657
              _pred->set(w,e);
658 658
            }
659 659
          }
660 660
          break;
661 661
        case Heap::POST_HEAP:
662 662
          break;
663 663
        }
664 664
      }
665 665
      return v;
666 666
    }
667 667

	
668 668
    ///The next node to be processed.
669 669

	
670 670
    ///Returns the next node to be processed or \c INVALID if the
671 671
    ///priority heap is empty.
672 672
    Node nextNode() const
673 673
    {
674 674
      return !_heap->empty()?_heap->top():INVALID;
675 675
    }
676 676

	
677 677
    ///Returns \c false if there are nodes to be processed.
678 678

	
679 679
    ///Returns \c false if there are nodes to be processed
680 680
    ///in the priority heap.
681 681
    bool emptyQueue() const { return _heap->empty(); }
682 682

	
683 683
    ///Returns the number of the nodes to be processed.
684 684

	
685 685
    ///Returns the number of the nodes to be processed
686 686
    ///in the priority heap.
687 687
    int queueSize() const { return _heap->size(); }
688 688

	
689 689
    ///Executes the algorithm.
Ignore white space 6 line context
1 1
/* -*- mode: C++; indent-tabs-mode: nil; -*-
2 2
 *
3 3
 * This file is a part of LEMON, a generic C++ optimization library.
4 4
 *
5 5
 * Copyright (C) 2003-2009
6 6
 * Egervary Jeno Kombinatorikus Optimalizalasi Kutatocsoport
7 7
 * (Egervary Research Group on Combinatorial Optimization, EGRES).
8 8
 *
9 9
 * Permission to use, modify and distribute this software is granted
10 10
 * provided that this copyright notice appears in all copies. For
11 11
 * precise terms see the accompanying LICENSE file.
12 12
 *
13 13
 * This software is provided "AS IS" with no warranty of any kind,
14 14
 * express or implied, and with no claim as to its suitability for any
15 15
 * purpose.
16 16
 *
17 17
 */
18 18

	
19 19
#ifndef LEMON_DIM2_H
20 20
#define LEMON_DIM2_H
21 21

	
22 22
#include <iostream>
23 23

	
24
///\ingroup misc
24
///\ingroup geomdat
25 25
///\file
26 26
///\brief A simple two dimensional vector and a bounding box implementation
27
///
28
/// The class \ref lemon::dim2::Point "dim2::Point" implements
29
/// a two dimensional vector with the usual operations.
30
///
31
/// The class \ref lemon::dim2::Box "dim2::Box" can be used to determine
32
/// the rectangular bounding box of a set of
33
/// \ref lemon::dim2::Point "dim2::Point"'s.
34 27

	
35 28
namespace lemon {
36 29

	
37 30
  ///Tools for handling two dimensional coordinates
38 31

	
39 32
  ///This namespace is a storage of several
40 33
  ///tools for handling two dimensional coordinates
41 34
  namespace dim2 {
42 35

	
43
  /// \addtogroup misc
36
  /// \addtogroup geomdat
44 37
  /// @{
45 38

	
46 39
  /// Two dimensional vector (plain vector)
47 40

	
48 41
  /// A simple two dimensional vector (plain vector) implementation
49 42
  /// with the usual vector operations.
50 43
  template<typename T>
51 44
    class Point {
52 45

	
53 46
    public:
54 47

	
55 48
      typedef T Value;
56 49

	
57 50
      ///First coordinate
58 51
      T x;
59 52
      ///Second coordinate
60 53
      T y;
61 54

	
62 55
      ///Default constructor
63 56
      Point() {}
64 57

	
65 58
      ///Construct an instance from coordinates
66 59
      Point(T a, T b) : x(a), y(b) { }
67 60

	
68 61
      ///Returns the dimension of the vector (i.e. returns 2).
69 62

	
70 63
      ///The dimension of the vector.
71 64
      ///This function always returns 2.
72 65
      int size() const { return 2; }
73 66

	
74 67
      ///Subscripting operator
75 68

	
76 69
      ///\c p[0] is \c p.x and \c p[1] is \c p.y
77 70
      ///
78 71
      T& operator[](int idx) { return idx == 0 ? x : y; }
79 72

	
80 73
      ///Const subscripting operator
81 74

	
82 75
      ///\c p[0] is \c p.x and \c p[1] is \c p.y
83 76
      ///
84 77
      const T& operator[](int idx) const { return idx == 0 ? x : y; }
85 78

	
86 79
      ///Conversion constructor
87 80
      template<class TT> Point(const Point<TT> &p) : x(p.x), y(p.y) {}
88 81

	
89 82
      ///Give back the square of the norm of the vector
90 83
      T normSquare() const {
91 84
        return x*x+y*y;
92 85
      }
93 86

	
94 87
      ///Increment the left hand side by \c u
95 88
      Point<T>& operator +=(const Point<T>& u) {
96 89
        x += u.x;
97 90
        y += u.y;
98 91
        return *this;
99 92
      }
100 93

	
101 94
      ///Decrement the left hand side by \c u
102 95
      Point<T>& operator -=(const Point<T>& u) {
103 96
        x -= u.x;
104 97
        y -= u.y;
105 98
        return *this;
106 99
      }
107 100

	
108 101
      ///Multiply the left hand side with a scalar
109 102
      Point<T>& operator *=(const T &u) {
110 103
        x *= u;
111 104
        y *= u;
112 105
        return *this;
113 106
      }
114 107

	
115 108
      ///Divide the left hand side by a scalar
116 109
      Point<T>& operator /=(const T &u) {
117 110
        x /= u;
118 111
        y /= u;
119 112
        return *this;
120 113
      }
121 114

	
122 115
      ///Return the scalar product of two vectors
123 116
      T operator *(const Point<T>& u) const {
124 117
        return x*u.x+y*u.y;
125 118
      }
126 119

	
127 120
      ///Return the sum of two vectors
128 121
      Point<T> operator+(const Point<T> &u) const {
129 122
        Point<T> b=*this;
130 123
        return b+=u;
131 124
      }
132 125

	
133 126
      ///Return the negative of the vector
134 127
      Point<T> operator-() const {
135 128
        Point<T> b=*this;
136 129
        b.x=-b.x; b.y=-b.y;
137 130
        return b;
138 131
      }
139 132

	
Ignore white space 6 line context
... ...
@@ -266,293 +266,293 @@
266 266
      
267 267
      while (sn != tn) {
268 268
	if ((*_order)[sn] < (*_order)[tn]) {
269 269
	  if ((*_weight)[tn] <= value) value = (*_weight)[tn];
270 270
	  tn = (*_pred)[tn];
271 271
	} else {
272 272
	  if ((*_weight)[sn] <= value) value = (*_weight)[sn];
273 273
	  sn = (*_pred)[sn];
274 274
	}
275 275
      }
276 276
      return value;
277 277
    }
278 278

	
279 279
    /// \brief Return the minimum cut between two nodes
280 280
    ///
281 281
    /// This function returns the minimum cut between the nodes \c s and \c t
282 282
    /// in the \c cutMap parameter by setting the nodes in the component of
283 283
    /// \c s to \c true and the other nodes to \c false.
284 284
    ///
285 285
    /// For higher level interfaces see MinCutNodeIt and MinCutEdgeIt.
286 286
    ///
287 287
    /// \param s The base node.
288 288
    /// \param t The node you want to separate from node \c s.
289 289
    /// \param cutMap The cut will be returned in this map.
290 290
    /// It must be a \c bool (or convertible) \ref concepts::ReadWriteMap
291 291
    /// "ReadWriteMap" on the graph nodes.
292 292
    ///
293 293
    /// \return The value of the minimum cut between \c s and \c t.
294 294
    ///
295 295
    /// \pre \ref run() must be called before using this function.
296 296
    template <typename CutMap>
297 297
    Value minCutMap(const Node& s, ///< 
298 298
                    const Node& t,
299 299
                    ///< 
300 300
                    CutMap& cutMap
301 301
                    ///< 
302 302
                    ) const {
303 303
      Node sn = s, tn = t;
304 304
      bool s_root=false;
305 305
      Node rn = INVALID;
306 306
      Value value = std::numeric_limits<Value>::max();
307 307
      
308 308
      while (sn != tn) {
309 309
	if ((*_order)[sn] < (*_order)[tn]) {
310 310
	  if ((*_weight)[tn] <= value) {
311 311
	    rn = tn;
312 312
            s_root = false;
313 313
	    value = (*_weight)[tn];
314 314
	  }
315 315
	  tn = (*_pred)[tn];
316 316
	} else {
317 317
	  if ((*_weight)[sn] <= value) {
318 318
	    rn = sn;
319 319
            s_root = true;
320 320
	    value = (*_weight)[sn];
321 321
	  }
322 322
	  sn = (*_pred)[sn];
323 323
	}
324 324
      }
325 325

	
326 326
      typename Graph::template NodeMap<bool> reached(_graph, false);
327 327
      reached[_root] = true;
328 328
      cutMap.set(_root, !s_root);
329 329
      reached[rn] = true;
330 330
      cutMap.set(rn, s_root);
331 331

	
332 332
      std::vector<Node> st;
333 333
      for (NodeIt n(_graph); n != INVALID; ++n) {
334 334
	st.clear();
335 335
        Node nn = n;
336 336
	while (!reached[nn]) {
337 337
	  st.push_back(nn);
338 338
	  nn = (*_pred)[nn];
339 339
	}
340 340
	while (!st.empty()) {
341 341
	  cutMap.set(st.back(), cutMap[nn]);
342 342
	  st.pop_back();
343 343
	}
344 344
      }
345 345
      
346 346
      return value;
347 347
    }
348 348

	
349 349
    ///@}
350 350

	
351 351
    friend class MinCutNodeIt;
352 352

	
353 353
    /// Iterate on the nodes of a minimum cut
354 354
    
355 355
    /// This iterator class lists the nodes of a minimum cut found by
356 356
    /// GomoryHu. Before using it, you must allocate a GomoryHu class
357 357
    /// and call its \ref GomoryHu::run() "run()" method.
358 358
    ///
359 359
    /// This example counts the nodes in the minimum cut separating \c s from
360 360
    /// \c t.
361 361
    /// \code
362
    /// GomoruHu<Graph> gom(g, capacities);
362
    /// GomoryHu<Graph> gom(g, capacities);
363 363
    /// gom.run();
364 364
    /// int cnt=0;
365
    /// for(GomoruHu<Graph>::MinCutNodeIt n(gom,s,t); n!=INVALID; ++n) ++cnt;
365
    /// for(GomoryHu<Graph>::MinCutNodeIt n(gom,s,t); n!=INVALID; ++n) ++cnt;
366 366
    /// \endcode
367 367
    class MinCutNodeIt
368 368
    {
369 369
      bool _side;
370 370
      typename Graph::NodeIt _node_it;
371 371
      typename Graph::template NodeMap<bool> _cut;
372 372
    public:
373 373
      /// Constructor
374 374

	
375 375
      /// Constructor.
376 376
      ///
377 377
      MinCutNodeIt(GomoryHu const &gomory,
378 378
                   ///< The GomoryHu class. You must call its
379 379
                   ///  run() method
380 380
                   ///  before initializing this iterator.
381 381
                   const Node& s, ///< The base node.
382 382
                   const Node& t,
383 383
                   ///< The node you want to separate from node \c s.
384 384
                   bool side=true
385 385
                   ///< If it is \c true (default) then the iterator lists
386 386
                   ///  the nodes of the component containing \c s,
387 387
                   ///  otherwise it lists the other component.
388 388
                   /// \note As the minimum cut is not always unique,
389 389
                   /// \code
390 390
                   /// MinCutNodeIt(gomory, s, t, true);
391 391
                   /// \endcode
392 392
                   /// and
393 393
                   /// \code
394 394
                   /// MinCutNodeIt(gomory, t, s, false);
395 395
                   /// \endcode
396 396
                   /// does not necessarily give the same set of nodes.
397 397
                   /// However it is ensured that
398 398
                   /// \code
399 399
                   /// MinCutNodeIt(gomory, s, t, true);
400 400
                   /// \endcode
401 401
                   /// and
402 402
                   /// \code
403 403
                   /// MinCutNodeIt(gomory, s, t, false);
404 404
                   /// \endcode
405 405
                   /// together list each node exactly once.
406 406
                   )
407 407
        : _side(side), _cut(gomory._graph)
408 408
      {
409 409
        gomory.minCutMap(s,t,_cut);
410 410
        for(_node_it=typename Graph::NodeIt(gomory._graph);
411 411
            _node_it!=INVALID && _cut[_node_it]!=_side;
412 412
            ++_node_it) {}
413 413
      }
414 414
      /// Conversion to \c Node
415 415

	
416 416
      /// Conversion to \c Node.
417 417
      ///
418 418
      operator typename Graph::Node() const
419 419
      {
420 420
        return _node_it;
421 421
      }
422 422
      bool operator==(Invalid) { return _node_it==INVALID; }
423 423
      bool operator!=(Invalid) { return _node_it!=INVALID; }
424 424
      /// Next node
425 425

	
426 426
      /// Next node.
427 427
      ///
428 428
      MinCutNodeIt &operator++()
429 429
      {
430 430
        for(++_node_it;_node_it!=INVALID&&_cut[_node_it]!=_side;++_node_it) {}
431 431
        return *this;
432 432
      }
433 433
      /// Postfix incrementation
434 434

	
435 435
      /// Postfix incrementation.
436 436
      ///
437 437
      /// \warning This incrementation
438 438
      /// returns a \c Node, not a \c MinCutNodeIt, as one may
439 439
      /// expect.
440 440
      typename Graph::Node operator++(int)
441 441
      {
442 442
        typename Graph::Node n=*this;
443 443
        ++(*this);
444 444
        return n;
445 445
      }
446 446
    };
447 447
    
448 448
    friend class MinCutEdgeIt;
449 449
    
450 450
    /// Iterate on the edges of a minimum cut
451 451
    
452 452
    /// This iterator class lists the edges of a minimum cut found by
453 453
    /// GomoryHu. Before using it, you must allocate a GomoryHu class
454 454
    /// and call its \ref GomoryHu::run() "run()" method.
455 455
    ///
456 456
    /// This example computes the value of the minimum cut separating \c s from
457 457
    /// \c t.
458 458
    /// \code
459
    /// GomoruHu<Graph> gom(g, capacities);
459
    /// GomoryHu<Graph> gom(g, capacities);
460 460
    /// gom.run();
461 461
    /// int value=0;
462
    /// for(GomoruHu<Graph>::MinCutEdgeIt e(gom,s,t); e!=INVALID; ++e)
462
    /// for(GomoryHu<Graph>::MinCutEdgeIt e(gom,s,t); e!=INVALID; ++e)
463 463
    ///   value+=capacities[e];
464 464
    /// \endcode
465 465
    /// The result will be the same as the value returned by
466 466
    /// \ref GomoryHu::minCutValue() "gom.minCutValue(s,t)".
467 467
    class MinCutEdgeIt
468 468
    {
469 469
      bool _side;
470 470
      const Graph &_graph;
471 471
      typename Graph::NodeIt _node_it;
472 472
      typename Graph::OutArcIt _arc_it;
473 473
      typename Graph::template NodeMap<bool> _cut;
474 474
      void step()
475 475
      {
476 476
        ++_arc_it;
477 477
        while(_node_it!=INVALID && _arc_it==INVALID)
478 478
          {
479 479
            for(++_node_it;_node_it!=INVALID&&!_cut[_node_it];++_node_it) {}
480 480
            if(_node_it!=INVALID)
481 481
              _arc_it=typename Graph::OutArcIt(_graph,_node_it);
482 482
          }
483 483
      }
484 484
      
485 485
    public:
486 486
      /// Constructor
487 487

	
488 488
      /// Constructor.
489 489
      ///
490 490
      MinCutEdgeIt(GomoryHu const &gomory,
491 491
                   ///< The GomoryHu class. You must call its
492 492
                   ///  run() method
493 493
                   ///  before initializing this iterator.
494 494
                   const Node& s,  ///< The base node.
495 495
                   const Node& t,
496 496
                   ///< The node you want to separate from node \c s.
497 497
                   bool side=true
498 498
                   ///< If it is \c true (default) then the listed arcs
499 499
                   ///  will be oriented from the
500 500
                   ///  nodes of the component containing \c s,
501 501
                   ///  otherwise they will be oriented in the opposite
502 502
                   ///  direction.
503 503
                   )
504 504
        : _graph(gomory._graph), _cut(_graph)
505 505
      {
506 506
        gomory.minCutMap(s,t,_cut);
507 507
        if(!side)
508 508
          for(typename Graph::NodeIt n(_graph);n!=INVALID;++n)
509 509
            _cut[n]=!_cut[n];
510 510

	
511 511
        for(_node_it=typename Graph::NodeIt(_graph);
512 512
            _node_it!=INVALID && !_cut[_node_it];
513 513
            ++_node_it) {}
514 514
        _arc_it = _node_it!=INVALID ?
515 515
          typename Graph::OutArcIt(_graph,_node_it) : INVALID;
516 516
        while(_node_it!=INVALID && _arc_it == INVALID)
517 517
          {
518 518
            for(++_node_it; _node_it!=INVALID&&!_cut[_node_it]; ++_node_it) {}
519 519
            if(_node_it!=INVALID)
520 520
              _arc_it= typename Graph::OutArcIt(_graph,_node_it);
521 521
          }
522 522
        while(_arc_it!=INVALID && _cut[_graph.target(_arc_it)]) step();
523 523
      }
524 524
      /// Conversion to \c Arc
525 525

	
526 526
      /// Conversion to \c Arc.
527 527
      ///
528 528
      operator typename Graph::Arc() const
529 529
      {
530 530
        return _arc_it;
531 531
      }
532 532
      /// Conversion to \c Edge
533 533

	
534 534
      /// Conversion to \c Edge.
535 535
      ///
536 536
      operator typename Graph::Edge() const
537 537
      {
538 538
        return _arc_it;
539 539
      }
540 540
      bool operator==(Invalid) { return _node_it==INVALID; }
541 541
      bool operator!=(Invalid) { return _node_it!=INVALID; }
542 542
      /// Next edge
543 543

	
544 544
      /// Next edge.
545 545
      ///
546 546
      MinCutEdgeIt &operator++()
547 547
      {
548 548
        step();
549 549
        while(_arc_it!=INVALID && _cut[_graph.target(_arc_it)]) step();
550 550
        return *this;
551 551
      }
552 552
      /// Postfix incrementation
553 553
      
554 554
      /// Postfix incrementation.
555 555
      ///
556 556
      /// \warning This incrementation
557 557
      /// returns an \c Arc, not a \c MinCutEdgeIt, as one may expect.
558 558
      typename Graph::Arc operator++(int)

Changeset was too big and was cut off... Show full diff

0 comments (0 inline)