Estimating exchange curve for inventory management through evolutionary multi-objective optimization

Abstract

Inventory management involves trade-offs between conflicting objectives such as cost minimization and service level maximization. The trade-off analysis of cycle stock investment and workload, so called the exchange curve, possibly dates back to several decades ago. These analyses seldom formulated inventory trade-offs as a multi-objective optimization problem and their solution procedures were all based on single objective optimization. To our best knowledge, there do exist some studies that propose non-classical approach to multi-objective inventory management. However, some of the objectives in earlier studies are not conflicted each other such that the multi-objective models were not properly justified. In this paper, a bi-objective inventory management model without redundancy is discussed first. Then a solution procedure based on evolutionary multi-objective optimization is introduced to effectively solve the fixed order model. The results show that the intrinsic multi-objective approach can find efficient policies of order size and safety factor simultaneously without estimating shortage cost or service level. Moreover, a fitted exchange curve of cost and service is useful in determining the best customer service possible for the given investment in inventory management. The cost-service trade-off can be observed in a single run of an iterative computation, so that it is more appropriate for the practice of inventory management. Key words: inventory management, exchange curve, multi-objective optimization, particle swarm optimization.