lemon/hao_orlin.h
author alpar
Mon, 12 Feb 2007 17:54:36 +0000
changeset 2360 72c7075ad5ba
parent 2275 ff46747676ed
child 2376 0ed45a6c74b1
permissions -rw-r--r--
Lagrange relaxation based algorithm for the delay constrained least cost
path problem.
     1 /* -*- C++ -*-
     2  *
     3  * This file is a part of LEMON, a generic C++ optimization library
     4  *
     5  * Copyright (C) 2003-2006
     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_HAO_ORLIN_H
    20 #define LEMON_HAO_ORLIN_H
    21 
    22 #include <cassert>
    23  
    24 
    25 
    26 #include <vector>
    27 #include <queue>
    28 #include <list>
    29 #include <limits>
    30 
    31 #include <lemon/maps.h>
    32 #include <lemon/graph_utils.h>
    33 #include <lemon/graph_adaptor.h>
    34 #include <lemon/iterable_maps.h>
    35 
    36 /// \file
    37 /// \ingroup flowalgs
    38 /// \brief Implementation of the Hao-Orlin algorithm.
    39 ///
    40 /// Implementation of the HaoOrlin algorithms class for testing network 
    41 /// reliability.
    42 
    43 namespace lemon {
    44 
    45   /// \ingroup flowalgs
    46   ///
    47   /// \brief %Hao-Orlin algorithm to find a minimum cut in directed graphs.
    48   ///
    49   /// Hao-Orlin calculates a minimum cut in a directed graph 
    50   /// \f$ D=(V,A) \f$. It takes a fixed node \f$ source \in V \f$ and consists
    51   /// of two phases: in the first phase it determines a minimum cut
    52   /// with \f$ source \f$ on the source-side (i.e. a set \f$ X\subsetneq V \f$
    53   /// with \f$ source \in X \f$ and minimal out-degree) and in the
    54   /// second phase it determines a minimum cut with \f$ source \f$ on the
    55   /// sink-side (i.e. a set \f$ X\subsetneq V \f$ with \f$ source \notin X \f$
    56   /// and minimal out-degree). Obviously, the smaller of these two
    57   /// cuts will be a minimum cut of \f$ D \f$. The algorithm is a
    58   /// modified push-relabel preflow algorithm and our implementation
    59   /// calculates the minimum cut in \f$ O(n^3) \f$ time (we use the
    60   /// highest-label rule). The purpose of such an algorithm is testing
    61   /// network reliability. For an undirected graph with \f$ n \f$
    62   /// nodes and \f$ e \f$ edges you can use the algorithm of Nagamochi
    63   /// and Ibaraki which solves the undirected problem in 
    64   /// \f$ O(ne + n^2 \log(n)) \f$ time: it is implemented in the MinCut 
    65   /// algorithm
    66   /// class.
    67   ///
    68   /// \param _Graph is the graph type of the algorithm.
    69   /// \param _CapacityMap is an edge map of capacities which should
    70   /// be any numreric type. The default type is _Graph::EdgeMap<int>.
    71   /// \param _Tolerance is the handler of the inexact computation. The
    72   /// default type for this is Tolerance<typename CapacityMap::Value>.
    73   ///
    74   /// \author Attila Bernath and Balazs Dezso
    75 #ifdef DOXYGEN
    76   template <typename _Graph, typename _CapacityMap, typename _Tolerance>
    77 #else
    78   template <typename _Graph,
    79 	    typename _CapacityMap = typename _Graph::template EdgeMap<int>,
    80             typename _Tolerance = Tolerance<typename _CapacityMap::Value> >
    81 #endif
    82   class HaoOrlin {
    83   protected:
    84 
    85     typedef _Graph Graph;
    86     typedef _CapacityMap CapacityMap;
    87     typedef _Tolerance Tolerance;
    88 
    89     typedef typename CapacityMap::Value Value;
    90 
    91     
    92     typedef typename Graph::Node Node;
    93     typedef typename Graph::NodeIt NodeIt;
    94     typedef typename Graph::EdgeIt EdgeIt;
    95     typedef typename Graph::OutEdgeIt OutEdgeIt;
    96     typedef typename Graph::InEdgeIt InEdgeIt;
    97 
    98     const Graph* _graph;
    99 
   100     const CapacityMap* _capacity;
   101 
   102     typedef typename Graph::template EdgeMap<Value> FlowMap;
   103 
   104     FlowMap* _preflow;
   105 
   106     Node _source, _target;
   107     int _node_num;
   108 
   109     typedef ResGraphAdaptor<const Graph, Value, CapacityMap, 
   110                             FlowMap, Tolerance> OutResGraph;
   111     typedef typename OutResGraph::Edge OutResEdge;
   112     
   113     OutResGraph* _out_res_graph;
   114 
   115     typedef RevGraphAdaptor<const Graph> RevGraph;
   116     RevGraph* _rev_graph;
   117 
   118     typedef ResGraphAdaptor<const RevGraph, Value, CapacityMap, 
   119                             FlowMap, Tolerance> InResGraph;
   120     typedef typename InResGraph::Edge InResEdge;
   121     
   122     InResGraph* _in_res_graph;
   123 
   124     typedef IterableBoolMap<Graph, Node> WakeMap;
   125     WakeMap* _wake;
   126 
   127     typedef typename Graph::template NodeMap<int> DistMap;
   128     DistMap* _dist;  
   129     
   130     typedef typename Graph::template NodeMap<Value> ExcessMap;
   131     ExcessMap* _excess;
   132 
   133     typedef typename Graph::template NodeMap<bool> SourceSetMap;
   134     SourceSetMap* _source_set;
   135 
   136     std::vector<int> _level_size;
   137 
   138     int _highest_active;
   139     std::vector<std::list<Node> > _active_nodes;
   140 
   141     int _dormant_max;
   142     std::vector<std::list<Node> > _dormant;
   143 
   144 
   145     Value _min_cut;
   146 
   147     typedef typename Graph::template NodeMap<bool> MinCutMap;
   148     MinCutMap* _min_cut_map;
   149 
   150     Tolerance _tolerance;
   151 
   152   public: 
   153 
   154     /// \brief Constructor
   155     ///
   156     /// Constructor of the algorithm class. 
   157     HaoOrlin(const Graph& graph, const CapacityMap& capacity, 
   158              const Tolerance& tolerance = Tolerance()) :
   159       _graph(&graph), _capacity(&capacity), 
   160       _preflow(0), _source(), _target(), 
   161       _out_res_graph(0), _rev_graph(0), _in_res_graph(0),
   162       _wake(0),_dist(0), _excess(0), _source_set(0), 
   163       _highest_active(), _active_nodes(), _dormant_max(), _dormant(), 
   164       _min_cut(), _min_cut_map(0), _tolerance(tolerance) {}
   165 
   166     ~HaoOrlin() {
   167       if (_min_cut_map) {
   168         delete _min_cut_map;
   169       } 
   170       if (_in_res_graph) {
   171         delete _in_res_graph;
   172       }
   173       if (_rev_graph) {
   174         delete _rev_graph;
   175       }
   176       if (_out_res_graph) {
   177         delete _out_res_graph;
   178       }
   179       if (_source_set) {
   180         delete _source_set;
   181       }
   182       if (_excess) {
   183         delete _excess;
   184       }
   185       if (_dist) {
   186         delete _dist;
   187       }
   188       if (_wake) {
   189         delete _wake;
   190       }
   191       if (_preflow) {
   192         delete _preflow;
   193       }
   194     }
   195     
   196   private:
   197     
   198     template <typename ResGraph>
   199     void findMinCut(const Node& target, bool out, ResGraph& res_graph) {
   200       typedef typename Graph::Node Node;
   201       typedef typename ResGraph::Edge ResEdge;
   202       typedef typename ResGraph::OutEdgeIt ResOutEdgeIt;
   203 
   204       for (typename Graph::EdgeIt it(*_graph); it != INVALID; ++it) {
   205         (*_preflow)[it] = 0;      
   206       }
   207       for (NodeIt it(*_graph); it != INVALID; ++it) {
   208         (*_wake)[it] = true;
   209         (*_dist)[it] = 1;
   210         (*_excess)[it] = 0;
   211         (*_source_set)[it] = false;
   212       }
   213 
   214       _dormant[0].push_front(_source);
   215       (*_source_set)[_source] = true;
   216       _dormant_max = 0;
   217       (*_wake)[_source] = false;
   218 
   219       _level_size[0] = 1;
   220       _level_size[1] = _node_num - 1;
   221 
   222       _target = target;
   223       (*_dist)[target] = 0;
   224 
   225       for (ResOutEdgeIt it(res_graph, _source); it != INVALID; ++it) {
   226         Value delta = res_graph.rescap(it);
   227         (*_excess)[_source] -= delta;
   228         res_graph.augment(it, delta);
   229         Node a = res_graph.target(it);
   230         if ((*_excess)[a] == 0 && (*_wake)[a] && a != _target) {
   231           _active_nodes[(*_dist)[a]].push_front(a);
   232           if (_highest_active < (*_dist)[a]) {
   233             _highest_active = (*_dist)[a];
   234           }
   235         }
   236         (*_excess)[a] += delta;
   237       }
   238 
   239 
   240       do {
   241 	Node n;
   242 	while ((n = findActiveNode()) != INVALID) {
   243 	  for (ResOutEdgeIt e(res_graph, n); e != INVALID; ++e) {
   244             Node a = res_graph.target(e);
   245             if ((*_dist)[a] >= (*_dist)[n] || !(*_wake)[a]) continue;
   246 	    Value delta = res_graph.rescap(e);
   247 	    if (_tolerance.positive((*_excess)[n] - delta)) {
   248               (*_excess)[n] -= delta;
   249 	    } else {
   250 	      delta = (*_excess)[n];
   251               (*_excess)[n] = 0;
   252 	    }
   253 	    res_graph.augment(e, delta);
   254 	    if ((*_excess)[a] == 0 && a != _target) {
   255 	      _active_nodes[(*_dist)[a]].push_front(a);
   256 	    }
   257 	    (*_excess)[a] += delta;
   258             if ((*_excess)[n] == 0) break;
   259 	  }
   260 	  if ((*_excess)[n] != 0) {
   261 	    relabel(n, res_graph);
   262           }
   263 	}
   264 
   265 	Value current_value = cutValue(out);
   266  	if (_min_cut > current_value){
   267           if (out) {
   268             for (NodeIt it(*_graph); it != INVALID; ++it) {
   269               _min_cut_map->set(it, !(*_wake)[it]);
   270             } 
   271           } else {
   272             for (NodeIt it(*_graph); it != INVALID; ++it) {
   273               _min_cut_map->set(it, (*_wake)[it]);
   274             } 
   275           }
   276 
   277 	  _min_cut = current_value;
   278  	}
   279 
   280       } while (selectNewSink(res_graph));
   281     }
   282 
   283     template <typename ResGraph>
   284     void relabel(const Node& n, ResGraph& res_graph) {
   285       typedef typename ResGraph::OutEdgeIt ResOutEdgeIt;
   286 
   287       int k = (*_dist)[n];
   288       if (_level_size[k] == 1) {
   289 	++_dormant_max;
   290 	for (NodeIt it(*_graph); it != INVALID; ++it) {
   291 	  if ((*_wake)[it] && (*_dist)[it] >= k) {
   292 	    (*_wake)[it] = false;
   293 	    _dormant[_dormant_max].push_front(it);
   294 	    --_level_size[(*_dist)[it]];
   295 	  }
   296 	}
   297         --_highest_active;
   298       } else {	
   299         int new_dist = _node_num;
   300         for (ResOutEdgeIt e(res_graph, n); e != INVALID; ++e) {
   301           Node t = res_graph.target(e);
   302           if ((*_wake)[t] && new_dist > (*_dist)[t]) {
   303             new_dist = (*_dist)[t];
   304           }
   305         }
   306         if (new_dist == _node_num) {
   307 	  ++_dormant_max;
   308 	  (*_wake)[n] = false;
   309 	  _dormant[_dormant_max].push_front(n);
   310 	  --_level_size[(*_dist)[n]];
   311 	} else {	    
   312 	  --_level_size[(*_dist)[n]];
   313 	  (*_dist)[n] = new_dist + 1;
   314 	  _highest_active = (*_dist)[n];
   315 	  _active_nodes[_highest_active].push_front(n);
   316 	  ++_level_size[(*_dist)[n]];
   317 	}
   318       }
   319     }
   320 
   321     template <typename ResGraph>
   322     bool selectNewSink(ResGraph& res_graph) {
   323       typedef typename ResGraph::OutEdgeIt ResOutEdgeIt;
   324 
   325       Node old_target = _target;
   326       (*_wake)[_target] = false;
   327       --_level_size[(*_dist)[_target]];
   328       _dormant[0].push_front(_target);
   329       (*_source_set)[_target] = true;
   330       if ((int)_dormant[0].size() == _node_num){
   331         _dormant[0].clear();
   332 	return false;
   333       }
   334 
   335       bool wake_was_empty = false;
   336 
   337       if(_wake->trueNum() == 0) {
   338 	while (!_dormant[_dormant_max].empty()){
   339 	  (*_wake)[_dormant[_dormant_max].front()] = true;
   340 	  ++_level_size[(*_dist)[_dormant[_dormant_max].front()]];
   341 	  _dormant[_dormant_max].pop_front();
   342 	}
   343 	--_dormant_max;
   344 	wake_was_empty = true;
   345       }
   346 
   347       int min_dist = std::numeric_limits<int>::max();
   348       for (typename WakeMap::TrueIt it(*_wake); it != INVALID; ++it) {
   349 	if (min_dist > (*_dist)[it]){
   350 	  _target = it;
   351 	  min_dist = (*_dist)[it];
   352 	}
   353       }
   354 
   355       if (wake_was_empty){
   356 	for (typename WakeMap::TrueIt it(*_wake); it != INVALID; ++it) {
   357           if ((*_excess)[it] != 0 && it != _target) {
   358             _active_nodes[(*_dist)[it]].push_front(it);
   359             if (_highest_active < (*_dist)[it]) {
   360               _highest_active = (*_dist)[it];		    
   361             }
   362 	  }
   363 	}
   364       }
   365 
   366       Node n = old_target;
   367       for (ResOutEdgeIt e(res_graph, n); e != INVALID; ++e){
   368         Node a = res_graph.target(e);
   369 	if (!(*_wake)[a]) continue;
   370         Value delta = res_graph.rescap(e);
   371         res_graph.augment(e, delta);
   372         (*_excess)[n] -= delta;
   373         if ((*_excess)[a] == 0 && (*_wake)[a] && a != _target) {
   374           _active_nodes[(*_dist)[a]].push_front(a);
   375           if (_highest_active < (*_dist)[a]) {
   376             _highest_active = (*_dist)[a];
   377           }
   378         }
   379         (*_excess)[a] += delta;
   380       }
   381       
   382       return true;
   383     }
   384 
   385     Node findActiveNode() {
   386       while (_highest_active > 0 && _active_nodes[_highest_active].empty()){ 
   387 	--_highest_active;
   388       }
   389       if( _highest_active > 0) {
   390        	Node n = _active_nodes[_highest_active].front();
   391 	_active_nodes[_highest_active].pop_front();
   392 	return n;
   393       } else {
   394 	return INVALID;
   395       }
   396     }
   397 
   398     Value cutValue(bool out) {
   399       Value value = 0;
   400       if (out) {
   401         for (typename WakeMap::TrueIt it(*_wake); it != INVALID; ++it) {
   402           for (InEdgeIt e(*_graph, it); e != INVALID; ++e) {
   403             if (!(*_wake)[_graph->source(e)]){
   404               value += (*_capacity)[e];
   405             }	
   406           }
   407         }
   408       } else {
   409         for (typename WakeMap::TrueIt it(*_wake); it != INVALID; ++it) {
   410           for (OutEdgeIt e(*_graph, it); e != INVALID; ++e) {
   411             if (!(*_wake)[_graph->target(e)]){
   412               value += (*_capacity)[e];
   413             }	
   414           }
   415         }
   416       }
   417       return value;
   418     }
   419 
   420 
   421   public:
   422 
   423     /// \name Execution control
   424     /// The simplest way to execute the algorithm is to use
   425     /// one of the member functions called \c run(...).
   426     /// \n
   427     /// If you need more control on the execution,
   428     /// first you must call \ref init(), then the \ref calculateIn() or
   429     /// \ref calculateIn() functions.
   430 
   431     /// @{
   432 
   433     /// \brief Initializes the internal data structures.
   434     ///
   435     /// Initializes the internal data structures. It creates
   436     /// the maps, residual graph adaptors and some bucket structures
   437     /// for the algorithm. 
   438     void init() {
   439       init(NodeIt(*_graph));
   440     }
   441 
   442     /// \brief Initializes the internal data structures.
   443     ///
   444     /// Initializes the internal data structures. It creates
   445     /// the maps, residual graph adaptor and some bucket structures
   446     /// for the algorithm. Node \c source  is used as the push-relabel
   447     /// algorithm's source.
   448     void init(const Node& source) {
   449       _source = source;
   450       _node_num = countNodes(*_graph);
   451 
   452       _dormant.resize(_node_num);
   453       _level_size.resize(_node_num, 0);
   454       _active_nodes.resize(_node_num);
   455 
   456       if (!_preflow) {
   457         _preflow = new FlowMap(*_graph);
   458       }
   459       if (!_wake) {
   460         _wake = new WakeMap(*_graph);
   461       }
   462       if (!_dist) {
   463         _dist = new DistMap(*_graph);
   464       }
   465       if (!_excess) {
   466         _excess = new ExcessMap(*_graph);
   467       }
   468       if (!_source_set) {
   469         _source_set = new SourceSetMap(*_graph);
   470       }
   471       if (!_out_res_graph) {
   472         _out_res_graph = new OutResGraph(*_graph, *_capacity, *_preflow);
   473       }
   474       if (!_rev_graph) {
   475         _rev_graph = new RevGraph(*_graph);
   476       }
   477       if (!_in_res_graph) {
   478         _in_res_graph = new InResGraph(*_rev_graph, *_capacity, *_preflow);
   479       }
   480       if (!_min_cut_map) {
   481         _min_cut_map = new MinCutMap(*_graph);
   482       }
   483 
   484       _min_cut = std::numeric_limits<Value>::max();
   485     }
   486 
   487 
   488     /// \brief Calculates a minimum cut with \f$ source \f$ on the
   489     /// source-side.
   490     ///
   491     /// \brief Calculates a minimum cut with \f$ source \f$ on the
   492     /// source-side (i.e. a set \f$ X\subsetneq V \f$ with \f$ source \in X \f$
   493     ///  and minimal out-degree).
   494     void calculateOut() {
   495       for (NodeIt it(*_graph); it != INVALID; ++it) {
   496         if (it != _source) {
   497           calculateOut(it);
   498           return;
   499         }
   500       }
   501     }
   502 
   503     /// \brief Calculates a minimum cut with \f$ source \f$ on the
   504     /// source-side.
   505     ///
   506     /// \brief Calculates a minimum cut with \f$ source \f$ on the
   507     /// source-side (i.e. a set \f$ X\subsetneq V \f$ with \f$ source \in X \f$
   508     ///  and minimal out-degree). The \c target is the initial target
   509     /// for the push-relabel algorithm.
   510     void calculateOut(const Node& target) {
   511       findMinCut(target, true, *_out_res_graph);
   512     }
   513 
   514     /// \brief Calculates a minimum cut with \f$ source \f$ on the
   515     /// sink-side.
   516     ///
   517     /// \brief Calculates a minimum cut with \f$ source \f$ on the
   518     /// sink-side (i.e. a set \f$ X\subsetneq V \f$ with 
   519     /// \f$ source \notin X \f$
   520     /// and minimal out-degree).
   521     void calculateIn() {
   522       for (NodeIt it(*_graph); it != INVALID; ++it) {
   523         if (it != _source) {
   524           calculateIn(it);
   525           return;
   526         }
   527       }
   528     }
   529 
   530     /// \brief Calculates a minimum cut with \f$ source \f$ on the
   531     /// sink-side.
   532     ///
   533     /// \brief Calculates a minimum cut with \f$ source \f$ on the
   534     /// sink-side (i.e. a set \f$ X\subsetneq V 
   535     /// \f$ with \f$ source \notin X \f$ and minimal out-degree).  
   536     /// The \c target is the initial
   537     /// target for the push-relabel algorithm.
   538     void calculateIn(const Node& target) {
   539       findMinCut(target, false, *_in_res_graph);
   540     }
   541 
   542     /// \brief Runs the algorithm.
   543     ///
   544     /// Runs the algorithm. It finds nodes \c source and \c target
   545     /// arbitrarily and then calls \ref init(), \ref calculateOut()
   546     /// and \ref calculateIn().
   547     void run() {
   548       init();
   549       for (NodeIt it(*_graph); it != INVALID; ++it) {
   550         if (it != _source) {
   551           calculateOut(it);
   552           calculateIn(it);
   553           return;
   554         }
   555       }
   556     }
   557 
   558     /// \brief Runs the algorithm.
   559     ///
   560     /// Runs the algorithm. It uses the given \c source node, finds a
   561     /// proper \c target and then calls the \ref init(), \ref
   562     /// calculateOut() and \ref calculateIn().
   563     void run(const Node& s) {
   564       init(s);
   565       for (NodeIt it(*_graph); it != INVALID; ++it) {
   566         if (it != _source) {
   567           calculateOut(it);
   568           calculateIn(it);
   569           return;
   570         }
   571       }
   572     }
   573 
   574     /// \brief Runs the algorithm.
   575     ///
   576     /// Runs the algorithm. It just calls the \ref init() and then
   577     /// \ref calculateOut() and \ref calculateIn().
   578     void run(const Node& s, const Node& t) {
   579       init(s); 
   580       calculateOut(t);
   581       calculateIn(t);
   582     }
   583 
   584     /// @}
   585     
   586     /// \name Query Functions 
   587     /// The result of the %HaoOrlin algorithm
   588     /// can be obtained using these functions.
   589     /// \n
   590     /// Before using these functions, either \ref run(), \ref
   591     /// calculateOut() or \ref calculateIn() must be called.
   592     
   593     /// @{
   594 
   595     /// \brief Returns the value of the minimum value cut.
   596     /// 
   597     /// Returns the value of the minimum value cut.
   598     Value minCut() const {
   599       return _min_cut;
   600     }
   601 
   602 
   603     /// \brief Returns a minimum cut.
   604     ///
   605     /// Sets \c nodeMap to the characteristic vector of a minimum
   606     /// value cut: it will give a nonempty set \f$ X\subsetneq V \f$
   607     /// with minimal out-degree (i.e. \c nodeMap will be true exactly
   608     /// for the nodes of \f$ X \f$).  \pre nodeMap should be a
   609     /// bool-valued node-map.
   610     template <typename NodeMap>
   611     Value minCut(NodeMap& nodeMap) const {
   612       for (NodeIt it(*_graph); it != INVALID; ++it) {
   613 	nodeMap.set(it, (*_min_cut_map)[it]);
   614       }
   615       return minCut();
   616     }
   617 
   618     /// @}
   619     
   620   }; //class HaoOrlin 
   621 
   622 
   623 } //namespace lemon
   624 
   625 #endif //LEMON_HAO_ORLIN_H