Multicolumn-multicut cross decomposition for stochastic mixed-integer linear programming

摘要

This paper proposes a multicolumn-multicut cross decomposition method for stochastic mixed-integer linear programming problems. In this method, multiple columns or multiple cuts are added to the Dantzig-Wolfe restricted master problem or the Benders relaxed master problem, in one iteration. We demonstrate the advantage of the proposed method in comparison with single cut and multicut Benders decomposition methods and a single column single cut cross decomposition method, through the case study of a bio-product supply chain problem.