Analysis and improvements of a solving procedure for the integration of lot-sizing and scheduling decisions

This paper addresses the integration of single-level lot-sizing and scheduling decisions in job-shops systems. These two production management phases are often performed separately, thus generating feasibility issues at operational level. We have recently presented an improved integrated approach, combining a Tabu Search and a Lagrangian heuristic, to deal with this problem. The goal here is to analyze the different improvements and implemented ideas that have allowed constructing a strong and consistent solving procedure. Numerical results show the pertinence of the modifications performed on several parts of the algorithm. The solution quality is improved for all groups of tested problem instances.