lemon/suurballe.h
author ladanyi
Tue, 22 Nov 2005 14:35:33 +0000
changeset 1826 e07a7a8acd77
parent 1435 8e85e6bbefdf
child 1875 98698b69a902
permissions -rw-r--r--
clean-up
     1 /* -*- C++ -*-
     2  * lemon/suurballe.h - Part of LEMON, a generic C++ optimization library
     3  *
     4  * Copyright (C) 2005 Egervary Jeno Kombinatorikus Optimalizalasi Kutatocsoport
     5  * (Egervary Research Group on Combinatorial Optimization, EGRES).
     6  *
     7  * Permission to use, modify and distribute this software is granted
     8  * provided that this copyright notice appears in all copies. For
     9  * precise terms see the accompanying LICENSE file.
    10  *
    11  * This software is provided "AS IS" with no warranty of any kind,
    12  * express or implied, and with no claim as to its suitability for any
    13  * purpose.
    14  *
    15  */
    16 
    17 #ifndef LEMON_SUURBALLE_H
    18 #define LEMON_SUURBALLE_H
    19 
    20 ///\ingroup flowalgs
    21 ///\file
    22 ///\brief An algorithm for finding k paths of minimal total length.
    23 
    24 
    25 #include <lemon/maps.h>
    26 #include <vector>
    27 #include <lemon/min_cost_flow.h>
    28 
    29 namespace lemon {
    30 
    31 /// \addtogroup flowalgs
    32 /// @{
    33 
    34   ///\brief Implementation of an algorithm for finding k edge-disjoint paths between 2 nodes 
    35   /// of minimal total length 
    36   ///
    37   /// The class \ref lemon::Suurballe implements
    38   /// an algorithm for finding k edge-disjoint paths
    39   /// from a given source node to a given target node in an
    40   /// edge-weighted directed graph having minimal total weight (length).
    41   ///
    42   ///\warning Length values should be nonnegative!
    43   /// 
    44   ///\param Graph The directed graph type the algorithm runs on.
    45   ///\param LengthMap The type of the length map (values should be nonnegative).
    46   ///
    47   ///\note It it questionable whether it is correct to call this method after
    48   ///%Suurballe for it is just a special case of Edmonds' and Karp's algorithm
    49   ///for finding minimum cost flows. In fact, this implementation just
    50   ///wraps the MinCostFlow algorithms. The paper of both %Suurballe and
    51   ///Edmonds-Karp published in 1972, therefore it is possibly right to
    52   ///state that they are
    53   ///independent results. Most frequently this special case is referred as
    54   ///%Suurballe method in the literature, especially in communication
    55   ///network context.
    56   ///\author Attila Bernath
    57   template <typename Graph, typename LengthMap>
    58   class Suurballe{
    59 
    60 
    61     typedef typename LengthMap::Value Length;
    62     
    63     typedef typename Graph::Node Node;
    64     typedef typename Graph::NodeIt NodeIt;
    65     typedef typename Graph::Edge Edge;
    66     typedef typename Graph::OutEdgeIt OutEdgeIt;
    67     typedef typename Graph::template EdgeMap<int> EdgeIntMap;
    68 
    69     typedef ConstMap<Edge,int> ConstMap;
    70 
    71     const Graph& G;
    72 
    73     Node s;
    74     Node t;
    75 
    76     //Auxiliary variables
    77     //This is the capacity map for the mincostflow problem
    78     ConstMap const1map;
    79     //This MinCostFlow instance will actually solve the problem
    80     MinCostFlow<Graph, LengthMap, ConstMap> min_cost_flow;
    81 
    82     //Container to store found paths
    83     std::vector< std::vector<Edge> > paths;
    84 
    85   public :
    86 
    87 
    88     /*! \brief The constructor of the class.
    89     
    90     \param _G The directed graph the algorithm runs on. 
    91     \param _length The length (weight or cost) of the edges. 
    92     \param _s Source node.
    93     \param _t Target node.
    94     */
    95     Suurballe(Graph& _G, LengthMap& _length, Node _s, Node _t) : 
    96       G(_G), s(_s), t(_t), const1map(1), 
    97       min_cost_flow(_G, _length, const1map, _s, _t) { }
    98 
    99     ///Runs the algorithm.
   100 
   101     ///Runs the algorithm.
   102     ///Returns k if there are at least k edge-disjoint paths from s to t.
   103     ///Otherwise it returns the number of edge-disjoint paths found 
   104     ///from s to t.
   105     ///
   106     ///\param k How many paths are we looking for?
   107     ///
   108     int run(int k) {
   109       int i = min_cost_flow.run(k);
   110 
   111       //Let's find the paths
   112       //We put the paths into stl vectors (as an inner representation). 
   113       //In the meantime we lose the information stored in 'reversed'.
   114       //We suppose the lengths to be positive now.
   115 
   116       //We don't want to change the flow of min_cost_flow, so we make a copy
   117       //The name here suggests that the flow has only 0/1 values.
   118       EdgeIntMap reversed(G); 
   119 
   120       for(typename Graph::EdgeIt e(G); e!=INVALID; ++e) 
   121 	reversed[e] = min_cost_flow.getFlow()[e];
   122       
   123       paths.clear();
   124       paths.resize(k);
   125       for (int j=0; j<i; ++j){
   126 	Node n=s;
   127 
   128 	while (n!=t){
   129 
   130 	  OutEdgeIt e(G, n);
   131 	  
   132 	  while (!reversed[e]){
   133 	    ++e;
   134 	  }
   135 	  n = G.target(e);
   136 	  paths[j].push_back(e);
   137 	  reversed[e] = 1-reversed[e];
   138 	}
   139 	
   140       }
   141       return i;
   142     }
   143 
   144     
   145     ///Returns the total length of the paths.
   146     
   147     ///This function gives back the total length of the found paths.
   148     Length totalLength(){
   149       return min_cost_flow.totalLength();
   150     }
   151 
   152     ///Returns the found flow.
   153 
   154     ///This function returns a const reference to the EdgeMap \c flow.
   155     const EdgeIntMap &getFlow() const { return min_cost_flow.flow;}
   156 
   157     /// Returns the optimal dual solution
   158     
   159     ///This function returns a const reference to the NodeMap
   160     ///\c potential (the dual solution).
   161     const EdgeIntMap &getPotential() const { return min_cost_flow.potential;}
   162 
   163     ///Checks whether the complementary slackness holds.
   164 
   165     ///This function checks, whether the given solution is optimal.
   166     ///Currently this function only checks optimality,
   167     ///doesn't bother with feasibility.
   168     ///It is meant for testing purposes.
   169     bool checkComplementarySlackness(){
   170       return min_cost_flow.checkComplementarySlackness();
   171     }
   172 
   173     ///Read the found paths.
   174     
   175     ///This function gives back the \c j-th path in argument p.
   176     ///Assumes that \c run() has been run and nothing has changed since then.
   177     /// \warning It is assumed that \c p is constructed to
   178     ///be a path of graph \c G.
   179     ///If \c j is not less than the result of previous \c run,
   180     ///then the result here will be an empty path (\c j can be 0 as well).
   181     ///
   182     ///\param Path The type of the path structure to put the result to (must meet lemon path concept).
   183     ///\param p The path to put the result to.
   184     ///\param j Which path you want to get from the found paths (in a real application you would get the found paths iteratively).
   185     template<typename Path>
   186     void getPath(Path& p, size_t j){
   187 
   188       p.clear();
   189       if (j>paths.size()-1){
   190 	return;
   191       }
   192       typename Path::Builder B(p);
   193       for(typename std::vector<Edge>::iterator i=paths[j].begin(); 
   194 	  i!=paths[j].end(); ++i ){
   195 	B.pushBack(*i);
   196       }
   197 
   198       B.commit();
   199     }
   200 
   201   }; //class Suurballe
   202 
   203   ///@}
   204 
   205 } //namespace lemon
   206 
   207 #endif //LEMON_SUURBALLE_H