mascoptDev.algos.graph
Class NetworkDesignRelaxCplex

java.lang.Object
  |
  +--mascoptDev.algos.graph.NetworkDesignRelaxCplex

public class NetworkDesignRelaxCplex
extends Object

Provides an algorithm to write and solve (by using an algorithm defined by A. Laugier and S. Petat ) the network design problem. This linear program is consider with a arc-path formulation, we so have generated some paths by using the class PathGeneratortRandomWalk. It is clear that we so just consider a part of the problem, because the number of paths we must consider is exponential. The hashmap phiMap, which is public, contains the paths generated. (for each request, a set of paths is associated) We have as entry : one graph G, one request graph D and a boolean (which specify if we consider the constraint of capacities or not) --- defineND defines the network design problem linear program --- separation is a method to add the blossom constraints which are violated and which compute the solution of the ND problem in which we have added these constraints --- printResult write the result.


Field Summary
 ilog.cplex.IloCplex cplex
           
 double INFINI
           
 HashMap phiMap
           
 HashMap varpMap
           
 HashMap xMap
           
 
Constructor Summary
NetworkDesignRelaxCplex(Graph G, Graph D, boolean withcapcons)
           
 
Method Summary
 void defineND()
           
 void printResult()
           
 void separation()
           
 
Methods inherited from class java.lang.Object
equals, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait
 

Field Detail

INFINI

public double INFINI

cplex

public ilog.cplex.IloCplex cplex

phiMap

public HashMap phiMap

varpMap

public HashMap varpMap

xMap

public HashMap xMap
Constructor Detail

NetworkDesignRelaxCplex

public NetworkDesignRelaxCplex(Graph G,
                               Graph D,
                               boolean withcapcons)
Method Detail

defineND

public void defineND()

printResult

public void printResult()

separation

public void separation()