gravatar
kpeter (Peter Kovacs)
kpeter@inf.elte.hu
Remove unnecessary integer requirement in Suurballe (#323)
0 1 0
default
1 file changed with 2 insertions and 5 deletions:
↑ Collapse diff ↑
Ignore white space 384 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_SUURBALLE_H
20 20
#define LEMON_SUURBALLE_H
21 21

	
22 22
///\ingroup shortest_path
23 23
///\file
24 24
///\brief An algorithm for finding arc-disjoint paths between two
25 25
/// nodes having minimum total length.
26 26

	
27 27
#include <vector>
28 28
#include <limits>
29 29
#include <lemon/bin_heap.h>
30 30
#include <lemon/path.h>
31 31
#include <lemon/list_graph.h>
32 32
#include <lemon/maps.h>
33 33

	
34 34
namespace lemon {
35 35

	
36 36
  /// \addtogroup shortest_path
37 37
  /// @{
38 38

	
39 39
  /// \brief Algorithm for finding arc-disjoint paths between two nodes
40 40
  /// having minimum total length.
41 41
  ///
42 42
  /// \ref lemon::Suurballe "Suurballe" implements an algorithm for
43 43
  /// finding arc-disjoint paths having minimum total length (cost)
44 44
  /// from a given source node to a given target node in a digraph.
45 45
  ///
46 46
  /// Note that this problem is a special case of the \ref min_cost_flow
47 47
  /// "minimum cost flow problem". This implementation is actually an
48 48
  /// efficient specialized version of the \ref CapacityScaling
49 49
  /// "Successive Shortest Path" algorithm directly for this problem.
50 50
  /// Therefore this class provides query functions for flow values and
51 51
  /// node potentials (the dual solution) just like the minimum cost flow
52 52
  /// algorithms.
53 53
  ///
54 54
  /// \tparam GR The digraph type the algorithm runs on.
55 55
  /// \tparam LEN The type of the length map.
56 56
  /// The default value is <tt>GR::ArcMap<int></tt>.
57 57
  ///
58
  /// \warning Length values should be \e non-negative \e integers.
58
  /// \warning Length values should be \e non-negative.
59 59
  ///
60 60
  /// \note For finding node-disjoint paths this algorithm can be used
61 61
  /// along with the \ref SplitNodes adaptor.
62 62
#ifdef DOXYGEN
63 63
  template <typename GR, typename LEN>
64 64
#else
65 65
  template < typename GR,
66 66
             typename LEN = typename GR::template ArcMap<int> >
67 67
#endif
68 68
  class Suurballe
69 69
  {
70 70
    TEMPLATE_DIGRAPH_TYPEDEFS(GR);
71 71

	
72 72
    typedef ConstMap<Arc, int> ConstArcMap;
73 73
    typedef typename GR::template NodeMap<Arc> PredMap;
74 74

	
75 75
  public:
76 76

	
77 77
    /// The type of the digraph the algorithm runs on.
78 78
    typedef GR Digraph;
79 79
    /// The type of the length map.
80 80
    typedef LEN LengthMap;
81 81
    /// The type of the lengths.
82 82
    typedef typename LengthMap::Value Length;
83 83
#ifdef DOXYGEN
84 84
    /// The type of the flow map.
85 85
    typedef GR::ArcMap<int> FlowMap;
86 86
    /// The type of the potential map.
87 87
    typedef GR::NodeMap<Length> PotentialMap;
88 88
#else
89 89
    /// The type of the flow map.
90 90
    typedef typename Digraph::template ArcMap<int> FlowMap;
91 91
    /// The type of the potential map.
92 92
    typedef typename Digraph::template NodeMap<Length> PotentialMap;
93 93
#endif
94 94

	
95 95
    /// The type of the path structures.
96 96
    typedef SimplePath<GR> Path;
97 97

	
98 98
  private:
99 99

	
100 100
    // ResidualDijkstra is a special implementation of the
101 101
    // Dijkstra algorithm for finding shortest paths in the
102 102
    // residual network with respect to the reduced arc lengths
103 103
    // and modifying the node potentials according to the
104 104
    // distance of the nodes.
105 105
    class ResidualDijkstra
106 106
    {
107 107
      typedef typename Digraph::template NodeMap<int> HeapCrossRef;
108 108
      typedef BinHeap<Length, HeapCrossRef> Heap;
109 109

	
110 110
    private:
111 111

	
112 112
      // The digraph the algorithm runs on
113 113
      const Digraph &_graph;
114 114

	
115 115
      // The main maps
116 116
      const FlowMap &_flow;
117 117
      const LengthMap &_length;
118 118
      PotentialMap &_potential;
119 119

	
120 120
      // The distance map
121 121
      PotentialMap _dist;
122 122
      // The pred arc map
123 123
      PredMap &_pred;
124 124
      // The processed (i.e. permanently labeled) nodes
125 125
      std::vector<Node> _proc_nodes;
126 126

	
127 127
      Node _s;
128 128
      Node _t;
129 129

	
130 130
    public:
131 131

	
132 132
      /// Constructor.
133 133
      ResidualDijkstra( const Digraph &graph,
134 134
                        const FlowMap &flow,
135 135
                        const LengthMap &length,
136 136
                        PotentialMap &potential,
137 137
                        PredMap &pred,
138 138
                        Node s, Node t ) :
139 139
        _graph(graph), _flow(flow), _length(length), _potential(potential),
140 140
        _dist(graph), _pred(pred), _s(s), _t(t) {}
141 141

	
142 142
      /// \brief Run the algorithm. It returns \c true if a path is found
143 143
      /// from the source node to the target node.
144 144
      bool run() {
145 145
        HeapCrossRef heap_cross_ref(_graph, Heap::PRE_HEAP);
146 146
        Heap heap(heap_cross_ref);
147 147
        heap.push(_s, 0);
148 148
        _pred[_s] = INVALID;
149 149
        _proc_nodes.clear();
150 150

	
151 151
        // Process nodes
152 152
        while (!heap.empty() && heap.top() != _t) {
153 153
          Node u = heap.top(), v;
154 154
          Length d = heap.prio() + _potential[u], nd;
155 155
          _dist[u] = heap.prio();
156 156
          heap.pop();
157 157
          _proc_nodes.push_back(u);
158 158

	
159 159
          // Traverse outgoing arcs
160 160
          for (OutArcIt e(_graph, u); e != INVALID; ++e) {
161 161
            if (_flow[e] == 0) {
162 162
              v = _graph.target(e);
163 163
              switch(heap.state(v)) {
164 164
              case Heap::PRE_HEAP:
165 165
                heap.push(v, d + _length[e] - _potential[v]);
166 166
                _pred[v] = e;
167 167
                break;
168 168
              case Heap::IN_HEAP:
169 169
                nd = d + _length[e] - _potential[v];
170 170
                if (nd < heap[v]) {
171 171
                  heap.decrease(v, nd);
172 172
                  _pred[v] = e;
173 173
                }
174 174
                break;
175 175
              case Heap::POST_HEAP:
176 176
                break;
177 177
              }
178 178
            }
179 179
          }
180 180

	
181 181
          // Traverse incoming arcs
182 182
          for (InArcIt e(_graph, u); e != INVALID; ++e) {
183 183
            if (_flow[e] == 1) {
184 184
              v = _graph.source(e);
185 185
              switch(heap.state(v)) {
186 186
              case Heap::PRE_HEAP:
187 187
                heap.push(v, d - _length[e] - _potential[v]);
188 188
                _pred[v] = e;
189 189
                break;
190 190
              case Heap::IN_HEAP:
191 191
                nd = d - _length[e] - _potential[v];
192 192
                if (nd < heap[v]) {
193 193
                  heap.decrease(v, nd);
194 194
                  _pred[v] = e;
195 195
                }
196 196
                break;
197 197
              case Heap::POST_HEAP:
198 198
                break;
199 199
              }
200 200
            }
201 201
          }
202 202
        }
203 203
        if (heap.empty()) return false;
204 204

	
205 205
        // Update potentials of processed nodes
206 206
        Length t_dist = heap.prio();
207 207
        for (int i = 0; i < int(_proc_nodes.size()); ++i)
208 208
          _potential[_proc_nodes[i]] += _dist[_proc_nodes[i]] - t_dist;
209 209
        return true;
210 210
      }
211 211

	
212 212
    }; //class ResidualDijkstra
213 213

	
214 214
  private:
215 215

	
216 216
    // The digraph the algorithm runs on
217 217
    const Digraph &_graph;
218 218
    // The length map
219 219
    const LengthMap &_length;
220 220

	
221 221
    // Arc map of the current flow
222 222
    FlowMap *_flow;
223 223
    bool _local_flow;
224 224
    // Node map of the current potentials
225 225
    PotentialMap *_potential;
226 226
    bool _local_potential;
227 227

	
228 228
    // The source node
229 229
    Node _source;
230 230
    // The target node
231 231
    Node _target;
232 232

	
233 233
    // Container to store the found paths
234 234
    std::vector< SimplePath<Digraph> > paths;
235 235
    int _path_num;
236 236

	
237 237
    // The pred arc map
238 238
    PredMap _pred;
239 239
    // Implementation of the Dijkstra algorithm for finding augmenting
240 240
    // shortest paths in the residual network
241 241
    ResidualDijkstra *_dijkstra;
242 242

	
243 243
  public:
244 244

	
245 245
    /// \brief Constructor.
246 246
    ///
247 247
    /// Constructor.
248 248
    ///
249 249
    /// \param graph The digraph the algorithm runs on.
250 250
    /// \param length The length (cost) values of the arcs.
251 251
    Suurballe( const Digraph &graph,
252 252
               const LengthMap &length ) :
253 253
      _graph(graph), _length(length), _flow(0), _local_flow(false),
254 254
      _potential(0), _local_potential(false), _pred(graph)
255
    {
256
      LEMON_ASSERT(std::numeric_limits<Length>::is_integer,
257
        "The length type of Suurballe must be integer");
258
    }
255
    {}
259 256

	
260 257
    /// Destructor.
261 258
    ~Suurballe() {
262 259
      if (_local_flow) delete _flow;
263 260
      if (_local_potential) delete _potential;
264 261
      delete _dijkstra;
265 262
    }
266 263

	
267 264
    /// \brief Set the flow map.
268 265
    ///
269 266
    /// This function sets the flow map.
270 267
    /// If it is not used before calling \ref run() or \ref init(),
271 268
    /// an instance will be allocated automatically. The destructor
272 269
    /// deallocates this automatically allocated map, of course.
273 270
    ///
274 271
    /// The found flow contains only 0 and 1 values, since it is the
275 272
    /// union of the found arc-disjoint paths.
276 273
    ///
277 274
    /// \return <tt>(*this)</tt>
278 275
    Suurballe& flowMap(FlowMap &map) {
279 276
      if (_local_flow) {
280 277
        delete _flow;
281 278
        _local_flow = false;
282 279
      }
283 280
      _flow = &map;
284 281
      return *this;
285 282
    }
286 283

	
287 284
    /// \brief Set the potential map.
288 285
    ///
289 286
    /// This function sets the potential map.
290 287
    /// If it is not used before calling \ref run() or \ref init(),
291 288
    /// an instance will be allocated automatically. The destructor
292 289
    /// deallocates this automatically allocated map, of course.
293 290
    ///
294 291
    /// The node potentials provide the dual solution of the underlying
295 292
    /// \ref min_cost_flow "minimum cost flow problem".
296 293
    ///
297 294
    /// \return <tt>(*this)</tt>
298 295
    Suurballe& potentialMap(PotentialMap &map) {
299 296
      if (_local_potential) {
300 297
        delete _potential;
301 298
        _local_potential = false;
302 299
      }
303 300
      _potential = &map;
304 301
      return *this;
305 302
    }
306 303

	
307 304
    /// \name Execution Control
308 305
    /// The simplest way to execute the algorithm is to call the run()
309 306
    /// function.
310 307
    /// \n
311 308
    /// If you only need the flow that is the union of the found
312 309
    /// arc-disjoint paths, you may call init() and findFlow().
313 310

	
314 311
    /// @{
315 312

	
316 313
    /// \brief Run the algorithm.
317 314
    ///
318 315
    /// This function runs the algorithm.
319 316
    ///
320 317
    /// \param s The source node.
321 318
    /// \param t The target node.
322 319
    /// \param k The number of paths to be found.
323 320
    ///
324 321
    /// \return \c k if there are at least \c k arc-disjoint paths from
325 322
    /// \c s to \c t in the digraph. Otherwise it returns the number of
326 323
    /// arc-disjoint paths found.
327 324
    ///
328 325
    /// \note Apart from the return value, <tt>s.run(s, t, k)</tt> is
329 326
    /// just a shortcut of the following code.
330 327
    /// \code
331 328
    ///   s.init(s);
332 329
    ///   s.findFlow(t, k);
333 330
    ///   s.findPaths();
334 331
    /// \endcode
335 332
    int run(const Node& s, const Node& t, int k = 2) {
336 333
      init(s);
337 334
      findFlow(t, k);
338 335
      findPaths();
339 336
      return _path_num;
340 337
    }
341 338

	
342 339
    /// \brief Initialize the algorithm.
343 340
    ///
344 341
    /// This function initializes the algorithm.
345 342
    ///
346 343
    /// \param s The source node.
347 344
    void init(const Node& s) {
348 345
      _source = s;
349 346

	
350 347
      // Initialize maps
351 348
      if (!_flow) {
352 349
        _flow = new FlowMap(_graph);
353 350
        _local_flow = true;
354 351
      }
355 352
      if (!_potential) {
356 353
        _potential = new PotentialMap(_graph);
357 354
        _local_potential = true;
358 355
      }
359 356
      for (ArcIt e(_graph); e != INVALID; ++e) (*_flow)[e] = 0;
360 357
      for (NodeIt n(_graph); n != INVALID; ++n) (*_potential)[n] = 0;
361 358
    }
362 359

	
363 360
    /// \brief Execute the algorithm to find an optimal flow.
364 361
    ///
365 362
    /// This function executes the successive shortest path algorithm to
366 363
    /// find a minimum cost flow, which is the union of \c k (or less)
367 364
    /// arc-disjoint paths.
368 365
    ///
369 366
    /// \param t The target node.
370 367
    /// \param k The number of paths to be found.
371 368
    ///
372 369
    /// \return \c k if there are at least \c k arc-disjoint paths from
373 370
    /// the source node to the given node \c t in the digraph.
374 371
    /// Otherwise it returns the number of arc-disjoint paths found.
375 372
    ///
376 373
    /// \pre \ref init() must be called before using this function.
377 374
    int findFlow(const Node& t, int k = 2) {
378 375
      _target = t;
379 376
      _dijkstra =
380 377
        new ResidualDijkstra( _graph, *_flow, _length, *_potential, _pred,
381 378
                              _source, _target );
382 379

	
383 380
      // Find shortest paths
384 381
      _path_num = 0;
385 382
      while (_path_num < k) {
386 383
        // Run Dijkstra
387 384
        if (!_dijkstra->run()) break;
388 385
        ++_path_num;
389 386

	
390 387
        // Set the flow along the found shortest path
391 388
        Node u = _target;
392 389
        Arc e;
393 390
        while ((e = _pred[u]) != INVALID) {
394 391
          if (u == _graph.target(e)) {
395 392
            (*_flow)[e] = 1;
396 393
            u = _graph.source(e);
397 394
          } else {
398 395
            (*_flow)[e] = 0;
399 396
            u = _graph.target(e);
400 397
          }
401 398
        }
402 399
      }
403 400
      return _path_num;
404 401
    }
405 402

	
406 403
    /// \brief Compute the paths from the flow.
407 404
    ///
408 405
    /// This function computes the paths from the found minimum cost flow,
409 406
    /// which is the union of some arc-disjoint paths.
410 407
    ///
411 408
    /// \pre \ref init() and \ref findFlow() must be called before using
412 409
    /// this function.
413 410
    void findPaths() {
414 411
      FlowMap res_flow(_graph);
415 412
      for(ArcIt a(_graph); a != INVALID; ++a) res_flow[a] = (*_flow)[a];
416 413

	
417 414
      paths.clear();
418 415
      paths.resize(_path_num);
419 416
      for (int i = 0; i < _path_num; ++i) {
420 417
        Node n = _source;
421 418
        while (n != _target) {
422 419
          OutArcIt e(_graph, n);
423 420
          for ( ; res_flow[e] == 0; ++e) ;
424 421
          n = _graph.target(e);
425 422
          paths[i].addBack(e);
426 423
          res_flow[e] = 0;
427 424
        }
428 425
      }
429 426
    }
430 427

	
431 428
    /// @}
432 429

	
433 430
    /// \name Query Functions
434 431
    /// The results of the algorithm can be obtained using these
435 432
    /// functions.
436 433
    /// \n The algorithm should be executed before using them.
437 434

	
438 435
    /// @{
439 436

	
440 437
    /// \brief Return the total length of the found paths.
441 438
    ///
442 439
    /// This function returns the total length of the found paths, i.e.
443 440
    /// the total cost of the found flow.
444 441
    /// The complexity of the function is O(e).
445 442
    ///
446 443
    /// \pre \ref run() or \ref findFlow() must be called before using
447 444
    /// this function.
448 445
    Length totalLength() const {
449 446
      Length c = 0;
450 447
      for (ArcIt e(_graph); e != INVALID; ++e)
0 comments (0 inline)