Novel Pareto-based meta-heuristics for solving multi-objective multi-item capacitated lot-sizing problems

Abstract

Capacitated production lot-sizing problems (LSPs) are challenging problems to solve due to their combinatorial nature. We consider a multi-item capacitated LSP with setup times, safety stocks, and demand shortages plus lost sales and backorder considerations for various production methods (i.e., job shop, batch flow, or continuous flow among others). We use multi-objective mathematical programming to solve this problem with three conflicting objectives including: (i) minimizing the total production costs; (ii) leveling the production volume in different production periods; and (iii) producing a solution which is as close as possible to the just-in-time level. We also consider lost sales, backorders, safety stocks, storage space limitation, and capacity constraints. We propose two novel Pareto-based multi-objective meta-heuristic algorithms: multi-objective vibration damping optimization (MOVDO) and a multi-objective harmony search algorithm (MOHSA). We compare MOVDO and MOHSA with two well-known evolutionary algorithms called the non-dominated sorting genetic algorithm (NSGA-II) and multi-objective simulated annealing (MOSA) to demonstrate the efficiency and effectiveness of the proposed methods.