A hybrid multi-objective approach based on the genetic algorithm and neural network to design an incremental cellular manufacturing system q Javad Rezaeian Zeidi a,⇑ , Nikbakhsh Javadian a,1 , Reza Tavakkoli-Moghaddam b,2 , Fariborz Jolai b,2 a Department of Industrial Engineering, Faculty of Engineering, Mazandaran University of Science and Technology, P.O. Box 734, Babol, Iran b Department of Industrial Engineering, Faculty of Engineering, University of Tehran, P.O. Box 11365-4563, Tehran, Iran article info Article history: Received 28 August 2010 Received in revised form 22 February 2012 Accepted 19 August 2013 Available online xxxx Keywords: Incremental cell formation Multi-objective optimization Genetic algorithm Neural network abstract One important issue related to the implementation of cellular manufacturing systems (CMSs) is to decide whether to convert an existing job shop into a CMS comprehensively in a single run, or in stages incre- mentally by forming cells one after the other, taking the advantage of the experiences of implementation. This paper presents a new multi-objective nonlinear programming model in a dynamic environment. Fur- thermore, a novel hybrid multi-objective approach based on the genetic algorithm and artificial neural network is proposed to solve the presented model. From the computational analyses, the proposed algo- rithm is found much more efficient than the fast non-dominated sorting genetic algorithm (NSGA-II) in generating Pareto optimal fronts. Ó 2013 Elsevier Ltd. All rights reserved. 1. Introduction The cellular manufacturing (CM) problem has captured a great deal of attention of many manufacturers and researchers. CM is the implementation of group technology (GT) and a manufacturing philosophy, in which similar parts are identified and grouped into part families; meanwhile, machines are grouped into machine cells to take advantage of their similarities in manufacturing and design (Chung, Wu, & Chang, 2011). CM, which is a flexible manufacturing system (FMS), can respond to the increasingly competitive envi- ronment facing manufacturers. Specially, manufacturers need to quickly improve their efficiency, response time and quality, but with a minimum of upfront investment of the capital and time. One of the implied assumptions in the modeling and development of CMSs is that the product mix remains stable over the time and a major deterrent to implement CMSs is changing the layout by entering new demands or variability of them. The variety and the uncertainty of demand, variety of characteristics of the product and manufacturing process are the reasons that motivated the re- quest for more flexibility. In the recent decades, it has been tried to develop new layouts and new models of cell formation with more flexibility (Hamedi, Esmaeilian, Ismail, & Ariffin, 2012). A number of researchers have suggested incremental cell formation in the literature. Cell formation is one of the important issues of CMSs, in which similar parts are grouped in a family known as part families and re- quired machines to process parts are determined. Many models and solution approaches have been developed to deal with a cell formation problem (CFP), but virtually all of them look at CM in terms of the total number of products to be made and the total number of machines or machine types available (or needed), and then try to plan a conversion of the entire shop into cells, possibly keeping a remainder cell. In other words, planning the conversion of a job shop to CM is performed comprehensively (non-incremen- tal) rather than incrementally (Marsh, Shafer, & Meredith, 1999). In situation of this kind, all machines belonging to shops will move to cells in a period totally, as shown in Fig. 1 (Rezaeian, Javadian, Tavakkoli-Moghaddam, & Jolai, 2011). Wemmerlov and Johnson (2000) in a survey carried out on 126 cells in 46 plants mentioned that academic (and some practitioner) writers on cell formation often seem to perceive the problem as one, in which multiple cells emerge from a single analysis of the factory. The reality is that most cells in industries are created and implemented sequentially over time. Incremental cell forma- tion follows a sequential process of forming the cells proposed in the master plan. In this case, cells are implemented one-by one rather than all-at-once, in which a sample of this kind of cell for- mation is illustrated in Fig. 2. According to this arrangement, 0360-8352/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.cie.2013.08.015 q This manuscript was processed by Area Editor Gursel A. Suer. ⇑ Corresponding author. Tel.: +98 1112291205; fax: +98 1112290118. E-mail address: j_rezaeian@ustmb.ac.ir (J.R. Zeidi). 1 Tel.: +98 1112291205; fax: +98 1112290118. 2 Tel.: +98 2161113358; fax: +98 2166409348. Computers & Industrial Engineering xxx (2013) xxx–xxx Contents lists available at ScienceDirect Computers & Industrial Engineering journal homepage: www.elsevier.com/locate/caie Please cite this article in press as: Zeidi, J. R., et al. A hybrid multi-objective approach based on the genetic algorithm and neural network to design an incremental cellular manufacturing system. Computers & Industrial Engineering (2013), http://dx.doi.org/10.1016/j.cie.2013.08.015