Computers & Operations Research 36 (2009) 777 – 794 www.elsevier.com/locate/cor A multi-objective scatter search for a dynamic cell formation problem M. Aramoon Bajestani, M. Rabbani ∗ , A.R. Rahimi-Vahed, G. Baharian Khoshkhou Department of Industrial Engineering. University of Tehran, P.O. Box 11365, 4563 Tehran, Iran Available online 6 November 2007 Abstract Cellular manufacturing system—an important application of group technology (GT)—has been recognized as an effective way to enhance the productivity in a factory. Consequently, a multi-objective dynamic cell formation problem is presented in this paper, where the total cell load variation and sum of the miscellaneous costs (machine cost, inter-cell material handling cost, and machine relocation cost) are to be minimized simultaneously. Since this type of problem is NP-hard, a new multi-objective scatter search (MOSS) is designed for finding locally Pareto-optimal frontier. To demonstrate the efficiency of the proposed algorithm, MOSS is compared with two salient multi-objective genetic algorithms, i.e. SPEA-II and NSGA-II based on some comparison metrics and statistical approach. The computational results indicate the superiority of the proposed MOSS compared to these two genetic algorithms.  2007 Elsevier Ltd. All rights reserved. Keywords: Multi-objective cell formation problem; Dynamic cell formation; Multi-objective scatter search; Multi-objective genetic algorithms 1. Introduction Group technology (GT) is a manufacturing philosophy which identifies and assigns the parts into part families and the machines into cells by taking advantage of part similarity in processing and design functions. One specific application of GT is cellular Manufacturing (CM) which strives to make the small-to-medium-sized batches of a large variety of part types produced in the flow shop manner (McAuley [1], King [2], and Ham et al. [3]). The major benefits of CM have been reported in the literature as simplification and reduction in material handling, decreasing the work-in-process inventories, reduction in set-up time, increment in flexibility, better production control, and shorter lead time (Askin and Estrada [4]). Identification of similar parts (part families) and machine cells in the design of a CM system (CMS) is commonly referred to as cell formation. Numerous current cell formation problems have been developed for a single period planning horizon. These assume that product demand and mix are constant for the entire planning horizon. Product mix signifies a collection of parts to be manufactured and product demand is the quantity of each part type to be produced. Yet considering the competitive nature of the industry nowadays, an applicable approach should take the dynamic conditions into account. As an illustration, the planning horizon should comprise a set of periods having a deterministic product demand and mix. Furthermore, cell reconfiguration is another promising strategy to consider in order for the manufacturing system to ∗ Corresponding author. E-mail address: mrabani@ut.ac.ir (M. Rabbani). 0305-0548/$ - see front matter  2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.cor.2007.10.026