Local Cuts and Two-Period Convex Hull Closures for Big-Bucket Lot-Sizing Problems

We study the big-bucket capacitated lot sizing problem with setup times. We use the novel methodology of Akartunali et al. (2014) that exploits two-period relaxations of the formulation in order to generate inequalities that cut-off the optimal solution of the linear programming relaxation. Our approach applies column generation in an unconventional way, with the master problem being a distance minimizing formulation and the subproblems being combina-torial two-period relaxations of the original problem. We identify a lower bound of the dimensionality of the generated cuts and provide extensive computational experiments that show how the generated bounds compare with other state-of- the-art approaches. Our results show that, for certain classes of problems, the bound improvement is considerable