In this work the combined problem is considered, which solves simultaneously the lot sizing and the cutting stock problems. We study some properties of the matrix of constraints and how to factorize the base without losing sparsity in the simplex method context, by a static reordering of the basic columns. Numerical results simulating simplex iterations and verify the sparsity of the factorizations are presented. Numerical experiments had also proven the robustness of this strategy. We conclude that the approach of constructing of the static sparse base reordering leads to very good computational results for both: speed and robustness, in comparison with approaches which do not consider the sparse structure of the matrix of constraints.
linear programming; cutting stock; lot sizing
