O.B. Lupanov et al. (Eds.): SAGA 2005, LNCS 3777, pp. 213 – 227, 2005. © Springer-Verlag Berlin Heidelberg 2005 Solving a Dynamic Cell Formation Problem with Machine Cost and Alternative Process Plan by Memetic Algorithms Reza Tavakkoli-Moghaddam 1 , Nima Safaei 2 , and Masoud Babakhani 2 1 Department of Industrial Engineering, Faculty of Engineering, University of Tehran, P.O. Box: 11365/4563, Tehran, Iran tavakoli@ut.ac.ir 2 Department of Industrial Engineering, Iran University of Science and Technology, P.C. 16846/13114, Tehran, Iran nima.safaei@iust.ac.ir Abstract. In this paper, we present a new model of a cell formation problem (CFP) for a multi-period planning horizon where the product mix and demand are different in each period, but they are deterministic. As a consequence, the formed cells in the current period may be not optimal for the next period. This evolution results from reformulation of part families, manufacturing cells, and reconfiguration of the CFP as required. Reconfiguration consists of reforming part families, machine groups, and machine relocations. The objective of the model is to determine the optimal number of cells while minimizing the machine amortization/relocation costs as well as the inter-cell movements in each period. In the proposed model, parts have alternative process plans, operation sequence, and produce as batch. The machine capacity is also limited and machine duplication is allowed. The proposed model for real-world instances cannot be solved optimally within a reasonable amount of computational time. Thus, we propose an efficient memetic algorithm (MA) with a simulated annealing-based local search engine for solving the proposed model. This model is solved optimally by the Lingo software then the optimal solution is compared with the MA implementation. Keywords: Dynamic cell formation, Alternative process plan, Machine relocation, Memetic Algorithm. 1 Introduction In most industries, the production is dynamic. In other words, the planning horizon can be divided into periods, in which each period has different product mix and demand requirements. In such cases, we face with dynamic production. Note that in the dynamic condition, product mix, and/or demand in each period are different from other periods but it is deterministic (i.e., known as a prior). In a dynamic production condition, the best cell formation (CF) design for one period may not be an efficient design for subsequent periods. By rearranging the manufacturing cells, the CF can continue operating efficiently as the product mix and demand changes. However, it may require some of machines moved from one cell to another cell (i.e., machine