lemon/preflow.h
author deba
Sun, 25 Nov 2007 22:56:44 +0000
changeset 2521 05c0ba99cc27
parent 2514 57143c09dc20
child 2522 616c019215c4
permissions -rw-r--r--
Bugfix: using read-write map instead reference map
     1 /* -*- C++ -*-
     2  *
     3  * This file is a part of LEMON, a generic C++ optimization library
     4  *
     5  * Copyright (C) 2003-2007
     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_PREFLOW_H
    20 #define LEMON_PREFLOW_H
    21 
    22 #include <lemon/error.h>
    23 #include <lemon/bits/invalid.h>
    24 #include <lemon/tolerance.h>
    25 #include <lemon/maps.h>
    26 #include <lemon/graph_utils.h>
    27 #include <lemon/elevator.h>
    28 
    29 /// \file
    30 /// \ingroup max_flow
    31 /// \brief Implementation of the preflow algorithm.
    32 
    33 namespace lemon { 
    34   
    35   /// \brief Default traits class of Preflow class.
    36   ///
    37   /// Default traits class of Preflow class.
    38   /// \param _Graph Graph type.
    39   /// \param _CapacityMap Type of capacity map.
    40   template <typename _Graph, typename _CapacityMap>
    41   struct PreflowDefaultTraits {
    42 
    43     /// \brief The graph type the algorithm runs on. 
    44     typedef _Graph Graph;
    45 
    46     /// \brief The type of the map that stores the edge capacities.
    47     ///
    48     /// The type of the map that stores the edge capacities.
    49     /// It must meet the \ref concepts::ReadMap "ReadMap" concept.
    50     typedef _CapacityMap CapacityMap;
    51 
    52     /// \brief The type of the length of the edges.
    53     typedef typename CapacityMap::Value Value;
    54 
    55     /// \brief The map type that stores the flow values.
    56     ///
    57     /// The map type that stores the flow values. 
    58     /// It must meet the \ref concepts::ReadWriteMap "ReadWriteMap" concept.
    59     typedef typename Graph::template EdgeMap<Value> FlowMap;
    60 
    61     /// \brief Instantiates a FlowMap.
    62     ///
    63     /// This function instantiates a \ref FlowMap. 
    64     /// \param graph The graph, to which we would like to define the flow map.
    65     static FlowMap* createFlowMap(const Graph& graph) {
    66       return new FlowMap(graph);
    67     }
    68 
    69     /// \brief The eleavator type used by Preflow algorithm.
    70     /// 
    71     /// The elevator type used by Preflow algorithm.
    72     ///
    73     /// \sa Elevator
    74     /// \sa LinkedElevator
    75     typedef LinkedElevator<Graph, typename Graph::Node> Elevator;
    76     
    77     /// \brief Instantiates an Elevator.
    78     ///
    79     /// This function instantiates a \ref Elevator. 
    80     /// \param graph The graph, to which we would like to define the elevator.
    81     /// \param max_level The maximum level of the elevator.
    82     static Elevator* createElevator(const Graph& graph, int max_level) {
    83       return new Elevator(graph, max_level);
    84     }
    85 
    86     /// \brief The tolerance used by the algorithm
    87     ///
    88     /// The tolerance used by the algorithm to handle inexact computation.
    89     typedef Tolerance<Value> Tolerance;
    90 
    91   };
    92   
    93 
    94   /// \ingroup max_flow
    95   ///
    96   /// \brief %Preflow algorithms class.
    97   ///
    98   /// This class provides an implementation of the Goldberg's \e
    99   /// preflow \e algorithm producing a flow of maximum value in a
   100   /// directed graph. The preflow algorithms are the fastest known max
   101   /// flow algorithms. The current implementation use a mixture of the
   102   /// \e "highest label" and the \e "bound decrease" heuristics.
   103   /// The worst case time complexity of the algorithm is \f$O(n^3)\f$.
   104   ///
   105   /// The algorithm consists from two phases. After the first phase
   106   /// the maximal flow value and the minimum cut can be obtained. The
   107   /// second phase constructs the feasible maximum flow on each edge.
   108   ///
   109   /// \param _Graph The directed graph type the algorithm runs on.
   110   /// \param _CapacityMap The flow map type.
   111   /// \param _Traits Traits class to set various data types used by
   112   /// the algorithm.  The default traits class is \ref
   113   /// PreflowDefaultTraits.  See \ref PreflowDefaultTraits for the
   114   /// documentation of a %Preflow traits class. 
   115   ///
   116   ///\author Jacint Szabo and Balazs Dezso
   117 #ifdef DOXYGEN
   118   template <typename _Graph, typename _CapacityMap, typename _Traits>
   119 #else
   120   template <typename _Graph, 
   121 	    typename _CapacityMap = typename _Graph::template EdgeMap<int>,
   122 	    typename _Traits = PreflowDefaultTraits<_Graph, _CapacityMap> >
   123 #endif
   124   class Preflow {
   125   public:
   126     
   127     typedef _Traits Traits;
   128     typedef typename Traits::Graph Graph;
   129     typedef typename Traits::CapacityMap CapacityMap;
   130     typedef typename Traits::Value Value; 
   131 
   132     typedef typename Traits::FlowMap FlowMap;
   133     typedef typename Traits::Elevator Elevator;
   134     typedef typename Traits::Tolerance Tolerance;
   135 
   136     /// \brief \ref Exception for uninitialized parameters.
   137     ///
   138     /// This error represents problems in the initialization
   139     /// of the parameters of the algorithms.
   140     class UninitializedParameter : public lemon::UninitializedParameter {
   141     public:
   142       virtual const char* what() const throw() {
   143 	return "lemon::Preflow::UninitializedParameter";
   144       }
   145     };
   146 
   147   private:
   148 
   149     GRAPH_TYPEDEFS(typename Graph);
   150 
   151     const Graph& _graph;
   152     const CapacityMap* _capacity;
   153 
   154     int _node_num;
   155 
   156     Node _source, _target;
   157 
   158     FlowMap* _flow;
   159     bool _local_flow;
   160 
   161     Elevator* _level;
   162     bool _local_level;
   163 
   164     typedef typename Graph::template NodeMap<Value> ExcessMap;
   165     ExcessMap* _excess;
   166 
   167     Tolerance _tolerance;
   168 
   169     bool _phase;
   170 
   171     void createStructures() {
   172       _node_num = countNodes(_graph);
   173 
   174       if (!_flow) {
   175 	_flow = Traits::createFlowMap(_graph);
   176 	_local_flow = true;
   177       }
   178       if (!_level) {
   179 	_level = Traits::createElevator(_graph, _node_num);
   180 	_local_level = true;
   181       }
   182       if (!_excess) {
   183 	_excess = new ExcessMap(_graph);
   184       }
   185     }
   186 
   187     void destroyStructures() {
   188       if (_local_flow) {
   189 	delete _flow;
   190       }
   191       if (_local_level) {
   192 	delete _level;
   193       }
   194       if (_excess) {
   195 	delete _excess;
   196       }
   197     }
   198 
   199   public:
   200 
   201     typedef Preflow Create;
   202 
   203     ///\name Named template parameters
   204 
   205     ///@{
   206 
   207     template <typename _FlowMap>
   208     struct DefFlowMapTraits : public Traits {
   209       typedef _FlowMap FlowMap;
   210       static FlowMap *createFlowMap(const Graph&) {
   211 	throw UninitializedParameter();
   212       }
   213     };
   214 
   215     /// \brief \ref named-templ-param "Named parameter" for setting
   216     /// FlowMap type
   217     ///
   218     /// \ref named-templ-param "Named parameter" for setting FlowMap
   219     /// type
   220     template <typename _FlowMap>
   221     struct DefFlowMap 
   222       : public Preflow<Graph, CapacityMap, DefFlowMapTraits<_FlowMap> > {
   223       typedef Preflow<Graph, CapacityMap, DefFlowMapTraits<_FlowMap> > Create;
   224     };
   225 
   226     template <typename _Elevator>
   227     struct DefElevatorTraits : public Traits {
   228       typedef _Elevator Elevator;
   229       static Elevator *createElevator(const Graph&, int) {
   230 	throw UninitializedParameter();
   231       }
   232     };
   233 
   234     /// \brief \ref named-templ-param "Named parameter" for setting
   235     /// Elevator type
   236     ///
   237     /// \ref named-templ-param "Named parameter" for setting Elevator
   238     /// type
   239     template <typename _Elevator>
   240     struct DefElevator 
   241       : public Preflow<Graph, CapacityMap, DefElevatorTraits<_Elevator> > {
   242       typedef Preflow<Graph, CapacityMap, DefElevatorTraits<_Elevator> > Create;
   243     };
   244 
   245     template <typename _Elevator>
   246     struct DefStandardElevatorTraits : public Traits {
   247       typedef _Elevator Elevator;
   248       static Elevator *createElevator(const Graph& graph, int max_level) {
   249 	return new Elevator(graph, max_level);
   250       }
   251     };
   252 
   253     /// \brief \ref named-templ-param "Named parameter" for setting
   254     /// Elevator type
   255     ///
   256     /// \ref named-templ-param "Named parameter" for setting Elevator
   257     /// type. The Elevator should be standard constructor interface, ie.
   258     /// the graph and the maximum level should be passed to it.
   259     template <typename _Elevator>
   260     struct DefStandardElevator 
   261       : public Preflow<Graph, CapacityMap, 
   262 		       DefStandardElevatorTraits<_Elevator> > {
   263       typedef Preflow<Graph, CapacityMap, 
   264 		      DefStandardElevatorTraits<_Elevator> > Create;
   265     };    
   266 
   267     /// @}
   268 
   269     /// \brief The constructor of the class.
   270     ///
   271     /// The constructor of the class. 
   272     /// \param graph The directed graph the algorithm runs on. 
   273     /// \param capacity The capacity of the edges. 
   274     /// \param source The source node.
   275     /// \param target The target node.
   276     Preflow(const Graph& graph, const CapacityMap& capacity, 
   277 	       Node source, Node target) 
   278       : _graph(graph), _capacity(&capacity), 
   279 	_node_num(0), _source(source), _target(target), 
   280 	_flow(0), _local_flow(false),
   281 	_level(0), _local_level(false),
   282 	_excess(0), _tolerance(), _phase() {}
   283 
   284     /// \brief Destrcutor.
   285     ///
   286     /// Destructor.
   287     ~Preflow() {
   288       destroyStructures();
   289     }
   290 
   291     /// \brief Sets the capacity map.
   292     ///
   293     /// Sets the capacity map.
   294     /// \return \c (*this)
   295     Preflow& capacityMap(const CapacityMap& map) {
   296       _capacity = &map;
   297       return *this;
   298     }
   299 
   300     /// \brief Sets the flow map.
   301     ///
   302     /// Sets the flow map.
   303     /// \return \c (*this)
   304     Preflow& flowMap(FlowMap& map) {
   305       if (_local_flow) {
   306 	delete _flow;
   307 	_local_flow = false;
   308       }
   309       _flow = &map;
   310       return *this;
   311     }
   312 
   313     /// \brief Returns the flow map.
   314     ///
   315     /// \return The flow map.
   316     const FlowMap& flowMap() {
   317       return *_flow;
   318     }
   319 
   320     /// \brief Sets the elevator.
   321     ///
   322     /// Sets the elevator.
   323     /// \return \c (*this)
   324     Preflow& elevator(Elevator& elevator) {
   325       if (_local_level) {
   326 	delete _level;
   327 	_local_level = false;
   328       }
   329       _level = &elevator;
   330       return *this;
   331     }
   332 
   333     /// \brief Returns the elevator.
   334     ///
   335     /// \return The elevator.
   336     const Elevator& elevator() {
   337       return *_level;
   338     }
   339 
   340     /// \brief Sets the source node.
   341     ///
   342     /// Sets the source node.
   343     /// \return \c (*this)
   344     Preflow& source(const Node& node) {
   345       _source = node;
   346       return *this;
   347     }
   348 
   349     /// \brief Sets the target node.
   350     ///
   351     /// Sets the target node.
   352     /// \return \c (*this)
   353     Preflow& target(const Node& node) {
   354       _target = node;
   355       return *this;
   356     }
   357  
   358     /// \brief Sets the tolerance used by algorithm.
   359     ///
   360     /// Sets the tolerance used by algorithm.
   361     Preflow& tolerance(const Tolerance& tolerance) const {
   362       _tolerance = tolerance;
   363       return *this;
   364     } 
   365 
   366     /// \brief Returns the tolerance used by algorithm.
   367     ///
   368     /// Returns the tolerance used by algorithm.
   369     const Tolerance& tolerance() const {
   370       return tolerance;
   371     } 
   372 
   373     /// \name Execution control The simplest way to execute the
   374     /// algorithm is to use one of the member functions called \c
   375     /// run().  
   376     /// \n
   377     /// If you need more control on initial solution or
   378     /// execution then you have to call one \ref init() function and then
   379     /// the startFirstPhase() and if you need the startSecondPhase().  
   380     
   381     ///@{
   382 
   383     /// \brief Initializes the internal data structures.
   384     ///
   385     /// Initializes the internal data structures.
   386     ///
   387     void init() {
   388       createStructures();
   389 
   390       _phase = true;
   391       for (NodeIt n(_graph); n != INVALID; ++n) {
   392 	_excess->set(n, 0);
   393       }
   394 
   395       for (EdgeIt e(_graph); e != INVALID; ++e) {
   396 	_flow->set(e, 0);
   397       }
   398 
   399       typename Graph::template NodeMap<bool> reached(_graph, false);
   400 
   401       _level->initStart();
   402       _level->initAddItem(_target);
   403 
   404       std::vector<Node> queue;
   405       reached.set(_source, true);
   406 
   407       queue.push_back(_target);
   408       reached.set(_target, true);
   409       while (!queue.empty()) {
   410 	std::vector<Node> nqueue;
   411 	for (int i = 0; i < int(queue.size()); ++i) {
   412 	  Node n = queue[i];
   413 	  for (InEdgeIt e(_graph, n); e != INVALID; ++e) {
   414 	    Node u = _graph.source(e);
   415 	    if (!reached[u] && _tolerance.positive((*_capacity)[e])) {
   416 	      reached.set(u, true);
   417 	      _level->initAddItem(u);
   418 	      nqueue.push_back(u);
   419 	    }
   420 	  }
   421 	}
   422 	queue.swap(nqueue);
   423       }
   424       _level->initFinish();
   425 
   426       for (OutEdgeIt e(_graph, _source); e != INVALID; ++e) {
   427 	if (_tolerance.positive((*_capacity)[e])) {
   428 	  Node u = _graph.target(e);
   429 	  if ((*_level)[u] == _level->maxLevel()) continue;
   430 	  _flow->set(e, (*_capacity)[e]);
   431 	  _excess->set(u, (*_excess)[u] + (*_capacity)[e]);
   432 	  if (u != _target && !_level->active(u)) {
   433 	    _level->activate(u);
   434 	  }
   435 	}
   436       }
   437     }
   438 
   439     /// \brief Initializes the internal data structures.
   440     ///
   441     /// Initializes the internal data structures and sets the initial
   442     /// flow to the given \c flowMap. The \c flowMap should contain a
   443     /// flow or at least a preflow, ie. in each node excluding the
   444     /// target the incoming flow should greater or equal to the
   445     /// outgoing flow.
   446     /// \return %False when the given \c flowMap is not a preflow. 
   447     template <typename FlowMap>
   448     bool flowInit(const FlowMap& flowMap) {
   449       createStructures();
   450 
   451       for (EdgeIt e(_graph); e != INVALID; ++e) {
   452 	_flow->set(e, flowMap[e]);
   453       }
   454 
   455       for (NodeIt n(_graph); n != INVALID; ++n) {
   456 	Value excess = 0;
   457 	for (InEdgeIt e(_graph, n); e != INVALID; ++e) {
   458 	  excess += (*_flow)[e];
   459 	}
   460 	for (OutEdgeIt e(_graph, n); e != INVALID; ++e) {
   461 	  excess -= (*_flow)[e];
   462 	}
   463 	if (excess < 0 && n != _source) return false;
   464 	_excess->set(n, excess);
   465       }
   466 
   467       typename Graph::template NodeMap<bool> reached(_graph, false);
   468 
   469       _level->initStart();
   470       _level->initAddItem(_target);
   471 
   472       std::vector<Node> queue;
   473       reached.set(_source, true);
   474 
   475       queue.push_back(_target);
   476       reached.set(_target, true);
   477       while (!queue.empty()) {
   478 	_level->initNewLevel();
   479 	std::vector<Node> nqueue;
   480 	for (int i = 0; i < int(queue.size()); ++i) {
   481 	  Node n = queue[i];
   482 	  for (InEdgeIt e(_graph, n); e != INVALID; ++e) {
   483 	    Node u = _graph.source(e);
   484 	    if (!reached[u] && 
   485 		_tolerance.positive((*_capacity)[e] - (*_flow)[e])) {
   486 	      reached.set(u, true);
   487 	      _level->initAddItem(u);
   488 	      nqueue.push_back(u);
   489 	    }
   490 	  }
   491 	  for (OutEdgeIt e(_graph, n); e != INVALID; ++e) {
   492 	    Node v = _graph.target(e);
   493 	    if (!reached[v] && _tolerance.positive((*_flow)[e])) {
   494 	      reached.set(v, true);
   495 	      _level->initAddItem(v);
   496 	      nqueue.push_back(v);
   497 	    }
   498 	  }
   499 	}
   500 	queue.swap(nqueue);
   501       }
   502       _level->initFinish();
   503 
   504       for (OutEdgeIt e(_graph, _source); e != INVALID; ++e) {
   505 	Value rem = (*_capacity)[e] - (*_flow)[e]; 
   506 	if (_tolerance.positive(rem)) {
   507 	  Node u = _graph.target(e);
   508 	  if ((*_level)[u] == _level->maxLevel()) continue;
   509 	  _flow->set(e, (*_capacity)[e]);
   510 	  _excess->set(u, (*_excess)[u] + rem);
   511 	  if (u != _target && !_level->active(u)) {
   512 	    _level->activate(u);
   513 	  }
   514 	}
   515       }
   516       for (InEdgeIt e(_graph, _source); e != INVALID; ++e) {
   517 	Value rem = (*_flow)[e]; 
   518 	if (_tolerance.positive(rem)) {
   519 	  Node v = _graph.source(e);
   520 	  if ((*_level)[v] == _level->maxLevel()) continue;
   521 	  _flow->set(e, 0);
   522 	  _excess->set(v, (*_excess)[v] + rem);
   523 	  if (v != _target && !_level->active(v)) {
   524 	    _level->activate(v);
   525 	  }
   526 	}
   527       }
   528       return true;
   529     }
   530 
   531     /// \brief Starts the first phase of the preflow algorithm.
   532     ///
   533     /// The preflow algorithm consists of two phases, this method runs
   534     /// the first phase. After the first phase the maximum flow value
   535     /// and a minimum value cut can already be computed, although a
   536     /// maximum flow is not yet obtained. So after calling this method
   537     /// \ref flowValue() returns the value of a maximum flow and \ref
   538     /// minCut() returns a minimum cut.     
   539     /// \pre One of the \ref init() functions should be called. 
   540     void startFirstPhase() {
   541       _phase = true;
   542 
   543       Node n = _level->highestActive();
   544       int level = _level->highestActiveLevel();
   545       while (n != INVALID) {
   546 	int num = _node_num;
   547 
   548 	while (num > 0 && n != INVALID) {
   549 	  Value excess = (*_excess)[n];
   550 	  int new_level = _level->maxLevel();
   551 
   552 	  for (OutEdgeIt e(_graph, n); e != INVALID; ++e) {
   553 	    Value rem = (*_capacity)[e] - (*_flow)[e];
   554 	    if (!_tolerance.positive(rem)) continue;
   555 	    Node v = _graph.target(e);
   556 	    if ((*_level)[v] < level) {
   557 	      if (!_level->active(v) && v != _target) {
   558 		_level->activate(v);
   559 	      }
   560 	      if (!_tolerance.less(rem, excess)) {
   561 		_flow->set(e, (*_flow)[e] + excess);
   562 		_excess->set(v, (*_excess)[v] + excess);
   563 		excess = 0;
   564 		goto no_more_push_1;
   565 	      } else {
   566 		excess -= rem;
   567 		_excess->set(v, (*_excess)[v] + rem);
   568 		_flow->set(e, (*_capacity)[e]);
   569 	      }
   570 	    } else if (new_level > (*_level)[v]) {
   571 	      new_level = (*_level)[v];
   572 	    }
   573 	  }
   574 
   575 	  for (InEdgeIt e(_graph, n); e != INVALID; ++e) {
   576 	    Value rem = (*_flow)[e];
   577 	    if (!_tolerance.positive(rem)) continue;
   578 	    Node v = _graph.source(e);
   579 	    if ((*_level)[v] < level) {
   580 	      if (!_level->active(v) && v != _target) {
   581 		_level->activate(v);
   582 	      }
   583 	      if (!_tolerance.less(rem, excess)) {
   584 		_flow->set(e, (*_flow)[e] - excess);
   585 		_excess->set(v, (*_excess)[v] + excess);
   586 		excess = 0;
   587 		goto no_more_push_1;
   588 	      } else {
   589 		excess -= rem;
   590 		_excess->set(v, (*_excess)[v] + rem);
   591 		_flow->set(e, 0);
   592 	      }
   593 	    } else if (new_level > (*_level)[v]) {
   594 	      new_level = (*_level)[v];
   595 	    }
   596 	  }
   597 
   598 	no_more_push_1:
   599 
   600 	  _excess->set(n, excess);
   601 
   602 	  if (excess != 0) {
   603 	    if (new_level + 1 < _level->maxLevel()) {
   604 	      _level->liftHighestActive(new_level + 1);
   605 	    } else {
   606 	      _level->liftHighestActiveToTop();
   607 	    }
   608 	    if (_level->emptyLevel(level)) {
   609 	      _level->liftToTop(level);
   610 	    }
   611 	  } else {
   612 	    _level->deactivate(n);
   613 	  }
   614 	  
   615 	  n = _level->highestActive();
   616 	  level = _level->highestActiveLevel();
   617 	  --num;
   618 	}
   619 	
   620 	num = _node_num * 20;
   621 	while (num > 0 && n != INVALID) {
   622 	  Value excess = (*_excess)[n];
   623 	  int new_level = _level->maxLevel();
   624 
   625 	  for (OutEdgeIt e(_graph, n); e != INVALID; ++e) {
   626 	    Value rem = (*_capacity)[e] - (*_flow)[e];
   627 	    if (!_tolerance.positive(rem)) continue;
   628 	    Node v = _graph.target(e);
   629 	    if ((*_level)[v] < level) {
   630 	      if (!_level->active(v) && v != _target) {
   631 		_level->activate(v);
   632 	      }
   633 	      if (!_tolerance.less(rem, excess)) {
   634 		_flow->set(e, (*_flow)[e] + excess);
   635 		_excess->set(v, (*_excess)[v] + excess);
   636 		excess = 0;
   637 		goto no_more_push_2;
   638 	      } else {
   639 		excess -= rem;
   640 		_excess->set(v, (*_excess)[v] + rem);
   641 		_flow->set(e, (*_capacity)[e]);
   642 	      }
   643 	    } else if (new_level > (*_level)[v]) {
   644 	      new_level = (*_level)[v];
   645 	    }
   646 	  }
   647 
   648 	  for (InEdgeIt e(_graph, n); e != INVALID; ++e) {
   649 	    Value rem = (*_flow)[e];
   650 	    if (!_tolerance.positive(rem)) continue;
   651 	    Node v = _graph.source(e);
   652 	    if ((*_level)[v] < level) {
   653 	      if (!_level->active(v) && v != _target) {
   654 		_level->activate(v);
   655 	      }
   656 	      if (!_tolerance.less(rem, excess)) {
   657 		_flow->set(e, (*_flow)[e] - excess);
   658 		_excess->set(v, (*_excess)[v] + excess);
   659 		excess = 0;
   660 		goto no_more_push_2;
   661 	      } else {
   662 		excess -= rem;
   663 		_excess->set(v, (*_excess)[v] + rem);
   664 		_flow->set(e, 0);
   665 	      }
   666 	    } else if (new_level > (*_level)[v]) {
   667 	      new_level = (*_level)[v];
   668 	    }
   669 	  }
   670 
   671 	no_more_push_2:
   672 
   673 	  _excess->set(n, excess);
   674 
   675 	  if (excess != 0) {
   676 	    if (new_level + 1 < _level->maxLevel()) {
   677 	      _level->liftActiveOn(level, new_level + 1);
   678 	    } else {
   679 	      _level->liftActiveToTop(level);
   680 	    }
   681 	    if (_level->emptyLevel(level)) {
   682 	      _level->liftToTop(level);
   683 	    }
   684 	  } else {
   685 	    _level->deactivate(n);
   686 	  }
   687 
   688 	  while (level >= 0 && _level->activeFree(level)) {
   689 	    --level;
   690 	  }
   691 	  if (level == -1) {
   692 	    n = _level->highestActive();
   693 	    level = _level->highestActiveLevel();
   694 	  } else {
   695 	    n = _level->activeOn(level);
   696 	  }
   697 	  --num;
   698 	}
   699       }
   700     }
   701 
   702     /// \brief Starts the second phase of the preflow algorithm.
   703     ///
   704     /// The preflow algorithm consists of two phases, this method runs
   705     /// the second phase. After calling \ref init() and \ref
   706     /// startFirstPhase() and then \ref startSecondPhase(), \ref
   707     /// flowMap() return a maximum flow, \ref flowValue() returns the
   708     /// value of a maximum flow, \ref minCut() returns a minimum cut
   709     /// \pre The \ref init() and startFirstPhase() functions should be
   710     /// called before.
   711     void startSecondPhase() {
   712       _phase = false;
   713 
   714       typename Graph::template NodeMap<bool> reached(_graph);
   715       for (NodeIt n(_graph); n != INVALID; ++n) {
   716 	reached.set(n, (*_level)[n] < _level->maxLevel());
   717       }
   718 
   719       _level->initStart();
   720       _level->initAddItem(_source);
   721  
   722       std::vector<Node> queue;
   723       queue.push_back(_source);
   724       reached.set(_source, true);
   725 
   726       while (!queue.empty()) {
   727 	_level->initNewLevel();
   728 	std::vector<Node> nqueue;
   729 	for (int i = 0; i < int(queue.size()); ++i) {
   730 	  Node n = queue[i];
   731 	  for (OutEdgeIt e(_graph, n); e != INVALID; ++e) {
   732 	    Node v = _graph.target(e);
   733 	    if (!reached[v] && _tolerance.positive((*_flow)[e])) {
   734 	      reached.set(v, true);
   735 	      _level->initAddItem(v);
   736 	      nqueue.push_back(v);
   737 	    }
   738 	  }
   739 	  for (InEdgeIt e(_graph, n); e != INVALID; ++e) {
   740 	    Node u = _graph.source(e);
   741 	    if (!reached[u] && 
   742 		_tolerance.positive((*_capacity)[e] - (*_flow)[e])) {
   743 	      reached.set(u, true);
   744 	      _level->initAddItem(u);
   745 	      nqueue.push_back(u);
   746 	    }
   747 	  }
   748 	}
   749 	queue.swap(nqueue);
   750       }
   751       _level->initFinish();
   752 
   753       for (NodeIt n(_graph); n != INVALID; ++n) {
   754 	if (!reached[n]) {
   755 	  _level->markToBottom(n);
   756 	} else if ((*_excess)[n] > 0 && _target != n) {
   757 	  _level->activate(n);
   758 	}
   759       }
   760 
   761       Node n;
   762       while ((n = _level->highestActive()) != INVALID) {
   763 	Value excess = (*_excess)[n];
   764 	int level = _level->highestActiveLevel();
   765 	int new_level = _level->maxLevel();
   766 
   767 	for (OutEdgeIt e(_graph, n); e != INVALID; ++e) {
   768 	  Value rem = (*_capacity)[e] - (*_flow)[e];
   769 	  if (!_tolerance.positive(rem)) continue;
   770 	  Node v = _graph.target(e);
   771 	  if ((*_level)[v] < level) {
   772 	    if (!_level->active(v) && v != _source) {
   773 	      _level->activate(v);
   774 	    }
   775 	    if (!_tolerance.less(rem, excess)) {
   776 	      _flow->set(e, (*_flow)[e] + excess);
   777 	      _excess->set(v, (*_excess)[v] + excess);
   778 	      excess = 0;
   779 	      goto no_more_push;
   780 	    } else {
   781 	      excess -= rem;
   782 	      _excess->set(v, (*_excess)[v] + rem);
   783 	      _flow->set(e, (*_capacity)[e]);
   784 	    }
   785 	  } else if (new_level > (*_level)[v]) {
   786 	    new_level = (*_level)[v];
   787 	  }
   788 	}
   789 
   790 	for (InEdgeIt e(_graph, n); e != INVALID; ++e) {
   791 	  Value rem = (*_flow)[e];
   792 	  if (!_tolerance.positive(rem)) continue;
   793 	  Node v = _graph.source(e);
   794 	  if ((*_level)[v] < level) {
   795 	    if (!_level->active(v) && v != _source) {
   796 	      _level->activate(v);
   797 	    }
   798 	    if (!_tolerance.less(rem, excess)) {
   799 	      _flow->set(e, (*_flow)[e] - excess);
   800 	      _excess->set(v, (*_excess)[v] + excess);
   801 	      excess = 0;
   802 	      goto no_more_push;
   803 	    } else {
   804 	      excess -= rem;
   805 	      _excess->set(v, (*_excess)[v] + rem);
   806 	      _flow->set(e, 0);
   807 	    }
   808 	  } else if (new_level > (*_level)[v]) {
   809 	    new_level = (*_level)[v];
   810 	  }
   811 	}
   812 
   813       no_more_push:
   814 
   815 	_excess->set(n, excess);
   816       
   817 	if (excess != 0) {
   818 	  if (new_level + 1 < _level->maxLevel()) {
   819 	    _level->liftHighestActive(new_level + 1);
   820 	  } else {
   821 	    // Calculation error 
   822 	    _level->liftHighestActiveToTop();
   823 	  }
   824 	  if (_level->emptyLevel(level)) {
   825 	    // Calculation error 
   826 	    _level->liftToTop(level);
   827 	  }
   828 	} else {
   829 	  _level->deactivate(n);
   830 	}
   831 
   832       }
   833     }
   834 
   835     /// \brief Runs the preflow algorithm.  
   836     ///
   837     /// Runs the preflow algorithm.
   838     /// \note pf.run() is just a shortcut of the following code.
   839     /// \code
   840     ///   pf.init();
   841     ///   pf.startFirstPhase();
   842     ///   pf.startSecondPhase();
   843     /// \endcode
   844     void run() {
   845       init();
   846       startFirstPhase();
   847       startSecondPhase();
   848     }
   849 
   850     /// \brief Runs the preflow algorithm to compute the minimum cut.  
   851     ///
   852     /// Runs the preflow algorithm to compute the minimum cut.
   853     /// \note pf.runMinCut() is just a shortcut of the following code.
   854     /// \code
   855     ///   pf.init();
   856     ///   pf.startFirstPhase();
   857     /// \endcode
   858     void runMinCut() {
   859       init();
   860       startFirstPhase();
   861     }
   862 
   863     /// @}
   864 
   865     /// \name Query Functions
   866     /// The result of the %Dijkstra algorithm can be obtained using these
   867     /// functions.\n
   868     /// Before the use of these functions,
   869     /// either run() or start() must be called.
   870     
   871     ///@{
   872 
   873     /// \brief Returns the value of the maximum flow.
   874     ///
   875     /// Returns the value of the maximum flow by returning the excess
   876     /// of the target node \c t. This value equals to the value of
   877     /// the maximum flow already after the first phase.
   878     Value flowValue() const {
   879       return (*_excess)[_target];
   880     }
   881 
   882     /// \brief Returns true when the node is on the source side of minimum cut.
   883     ///
   884     /// Returns true when the node is on the source side of minimum
   885     /// cut. This method can be called both after running \ref
   886     /// startFirstPhase() and \ref startSecondPhase().
   887     bool minCut(const Node& node) const {
   888       return ((*_level)[node] == _level->maxLevel()) == _phase;
   889     }
   890  
   891     /// \brief Returns a minimum value cut.
   892     ///
   893     /// Sets the \c cutMap to the characteristic vector of a minimum value
   894     /// cut. This method can be called both after running \ref
   895     /// startFirstPhase() and \ref startSecondPhase(). The result after second
   896     /// phase could be changed slightly if inexact computation is used.
   897     /// \pre The \c cutMap should be a bool-valued node-map.
   898     template <typename CutMap>
   899     void minCutMap(CutMap& cutMap) const {
   900       for (NodeIt n(_graph); n != INVALID; ++n) {
   901 	cutMap.set(n, minCut(n));
   902       }
   903     }
   904 
   905     /// \brief Returns the flow on the edge.
   906     ///
   907     /// Sets the \c flowMap to the flow on the edges. This method can
   908     /// be called after the second phase of algorithm.
   909     Value flow(const Edge& edge) const {
   910       return (*_flow)[edge];
   911     }
   912     
   913     /// @}    
   914   };
   915 }
   916 
   917 #endif