Title: A meta-heuristic for design of a capacitated production/inventory system

Authors: Cigdem Z. Gurgur

Addresses: Doermer School of Business and Management Sciences, Indiana – Purdue University, 2101 East Coliseum Blvd., Fort Wayne, IN 46805, USA

Abstract: In this paper, we address the optimal configuration of a capacitated production/inventory system with multiple products. The optimisation problem that we consider involves minimisation of the expected total inventory cost within the system subject to a service-level constraint. The functional relationship between the response variables and the system parameters cannot be expressed analytically. Moreover, the optimisation problem considered in this paper involves large number of decision variables. The size of the problem makes use of many numerical methods impractical. Therefore, as the optimisation tool, we utilise Response Surface Methodology, a heuristic sequential optimisation technique based on design of experiments and meta-modelling to approximate the functional relationship between the response variables and system parameters. Based on this model, we discuss the optimal design for a two-stage, three-product system.

Keywords: capacity planning; inventory management; multi-item; multi-echelon; multi-stage; decentralised production-inventory systems; pull type; numerical optimisation; stochastic modelling; metaheuristics; total inventory cost; service-level constraints; response surface methodology; RSM; metamodelling; capacitated production-inventory system.

DOI: 10.1504/IJSEM.2010.033375

International Journal of Services, Economics and Management, 2010 Vol.2 No.3/4, pp.412 - 432

Published online: 01 Jun 2010 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article