Incorporating facet-inducing inequalities into graphical-construct-based Lagrangian relaxation methodologies

This paper shows that facet-inducing inequalities can be incorporated into graphical-construct-based Lagrangian relaxation methodologies by including them as dualized constraints. It develops an algorithm that successively identifies additional facet-inducing inequalities and incorporates them into the Lagranian function. Computational experience shows that the algorithm can resolve duality gaps using relatively few facet-inducing inequalities.