Title: Dual-stage genetic algorithm approach for capacitated lot sizing problem

Authors: Ramadan Saleem; Gürsel A. Süer; Jing Huang

Addresses: Mechanical and Industrial Engineering Department, Applied Science Private University, P.O. Box 166, Amman, 11931, Jordan ' Industrial and Systems Engineering Department, Ohio University, Athens, OH 45701, USA ' Industrial and Systems Engineering Department, Ohio University, Athens, OH 45701, USA

Abstract: The capacitated lot sizing problem (CLSP) deals with determining lot sizes under capacity restrictions such that the total cost is minimised. The costs considered in this paper are ordering, inventory carrying, and labour costs. The work presented in this paper is an improvement over single stage genetic algorithm (SSGA) approach proposed by Süer et al. (2008) where no partial coverage was considered. This study addresses this weakness of their work and allows partial coverage by considering a dual-stage genetic algorithm (DSGA). Stage 1 of the proposed GA works with binary representation and the best chromosome is input into Stage 2. Stage 2 of the DSGA works with integer representation and uses only mutation operator. Stage 1 includes a group-based enumeration crossover strategy and Stage 2 includes a dynamic-percentage mutation strategy. The results of the experimentation showed that newly proposed DSGA outperformed SSGA in all test problems used in the experimentation.

Keywords: capacitated lot sizing; mathematical modelling; dual-stage genetic algorithms; ordering costs; inventory carrying costs; labour costs.

DOI: 10.1504/IJAOM.2013.058887

International Journal of Advanced Operations Management, 2013 Vol.5 No.4, pp.299 - 319

Received: 08 Sep 2012
Accepted: 28 Jan 2013

Published online: 27 Jan 2014 *

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