Hi all,
I am currently facing this issue: for modeling a sort of VRP problem I have the
need to find a solution and then perturb the problem to add a new decision
variable (say a new customer to be visited). I would like to let the search
engine explore the search space of the modified problem starting from the
solution of the original problem, is there a way to implement this behavior?
I make a concrete example, suppose the nodes to be visited are initially {1, 2,
3}, the decision variables is the set succ of successors (succ[i] \in {1..3}, i
= 1, …, 3) and the first solution satisfying also the other constraints is succ
= [3, 1, 2].
Now I need to add a node 4, so the successors becomes succ[i] \in {1..4}, i =1,
…, 4 and I would like to start the exploration of the tree close to the
solution succ = [3, 1, 2, _] (actually it should be [3, 1, 2, 4] because of
distinctness).
Thanks in advance for your answer.
All the best,
Luca
_______________________________________________
Gecode users mailing list
users@gecode.org
https://www.gecode.org/mailman/listinfo/gecode-users
