Relaxations for two-level multi-item lot-sizing problems (Q403668)

This article considers the generalized lot-sizing problem in combinatorial optimization. The article begins with a detailed overview of the existing literature and solution approaches to the classic lot-sizing problem including mixed integer formulations, before moving to problem extensions involving multiple levels, dependent demands and multiple items or clients. The focus of the research is to study the polyhedral structure of the optimization problem with a view to derive a strong formulation with valid inequalities which produce an efficient solution strategy. The optimization results are validated computationally on generated problem instances.