An Efficient Heuristic Algorithm for Capacitated Lot Sizing Problem with Overtime Decisions

Abstract

Capacitated Lot Sizing Problem is a very important tactical level decision making problem that answers the questions of producing when and how many in dynamic demand environment. Solving Capacitated Lot Sizing Problem with Overtime decisions (CLSPO) and extensions derived from the fundamental structure optimally suffer from combinatorial nature of the problem. The aim of the study is to form a two-stage heuristic algorithm to solve related problem in polynomial time. In first part, characteristics of problem structure are presented. Dominance properties are presented to help algorithm obtain a bounded search area. Proposed algorithm directly utilizes such shortcoming. Performance of approach is tested by using different criteria. And finally, robustness test are applied to check how well algorithm performs against fluctuations in its data. Simulated annealing as improvement heuristic performs well for related problem. It is also observed that fluctuations of data directly affects performance outcome. Obtained results also reveal that performance of improvement heuristic highly depends on constructive heuristic. Algorithm is also applied to an industry case study to plan master production schedule with minimum costs.