Variable Neighborhood Search Based Approach for Solving Multilevel Lot-Sizing Problems

In this paper, we present an effective approach based on the variable neighborhood search (VNS) for solving multilevel lot-sizing (MLLS) problems. Two kinds of neighborhood search strategies, i.e., move at first improvement (MAFI) and move at best improvement (MABI), are adopted to improve the performance of proposed algorithm. Computational experiments are carried out on 96 benchmark problems to test the optimality against genetic algorithm on identical problems, and also to analyze the mechanism of VNS while it solving MLLS problem. Experimental outcomes show that the VNS algorithm equipped with MABI and emendation by inner corner property enjoys good optimality and high computation effectiveness as well, which is quite competitive to the existing algorithms that have been studied on the MLLS problems. Key-Words: multilevel lot-sizing, variable neighborhood search, meta-heuristic, genetic algorithm