Adding Valid Inequalities within Bi-objective Binary Branch & Bound
Résumé
In order to enumerate every non-dominated solutions of a Bi-objective Binary linear program, we propose a Branch&Bound framework (BBBB). The efficiency of a BBBB is mainly based on the availability to prune the nodes using a comparison between lower bound and upper bound sets. In order to strengthen the lower bound sets, we add valid inequalities using classical separation methods coming from mono-objective optimization. In this bi-objective context, such cutting plane approach could be improved using multi-point separation algorithms. Using the VoptSolver framework in Julia, preliminary experimental results will be presented in order to show the efficiency of the method.