Heuristic family disaggregation techniques for hierarchical production planning systems

Abstract

Hierarchical production planning (HPP) systems are developed to establish a structured decision-making environment in large production systems which would otherwise become intractable. Essentially, in HPP systems the production policy for the aggregate level is determined and then disaggregated to the next level of product families. However, infeasibility issues always exist when imposing constraints from higher to lower levels. These issues are sometimes the offspring of the aggregate information used at higher levels. On the other hand, infeasibilities pertaining to the disaggregation algorithms also exist. Here, we resolve the algorithmic infeasibilities that exist in the family disaggregation procedure given by Bitran and Hax (1981) (BH). A heuristic modification (MBH) to procedure BH and an alternative filling procedure (FP) which eliminate backorders completely are introduced. The procedures reduce total set up costs. Empirical results demonstrating the performance of the heuristics are reported.