Originally posted by: Krzysztof Postek

A MILP is solved by means of B&B or CP, building a B&B tree, and solving an LP relaxation of the problem in each of the nodes in the tree. Is it possible to recover the dual optimal values for this LP relaxation? For each node, so far. Thanks in advance.

Hi,

could https://www.ibm.com/developerworks/community/forums/html/topic?id=77777777-0000-0000-0000-000014845703&ps=25

help ?

regards

Originally posted by: Krzysztof Postek

Hi,

thanks for the response, I have read that thread but it's not exactly what I look for since from their discussion it seems like you will often get zeros and you need to solve some fixed model.

I thought that, since the MILP is solved by a sequence of relaxations anyway, maybe for each of these relaxations it is possible just to extract the optimal dual vector? The dual information should be somewhere there since it's essentially an LP that is solved in each node.

Hi,

more information at http://www-01.ibm.com/support/docview.wss?uid=swg21399941

regards