A simulated annealing algorithm for dynamic system reconfiguration and production planning in cellular manufacturing

Most manufacturing system design problems have been studied under static conditions in which the facilities are configured on fixed shop floors for relatively long planning period by assuming constant product mix and demand. In today's dynamic business environment, shorter time periods should be considered where the product mix and demand may vary from period to period. As a result, the best facility layout for one period may not be efficient for subsequent periods. To address this issue, several authors proposed dynamic system reconfiguration models and solution procedures for manufacturing system design. In this paper, we consider an integrated problem of production planning and dynamic system reconfiguration in cellular manufacturing systems where production quantities are also decision variables. Based on this consideration, we propose a mathematical programming model for solving this problem. The solution of the model provides the planned production quantity, inventory level and system configuration for each period. Since the problem is NP-hard, we developed a heuristic algorithm based on multiple Markov chain simulated annealing to allow multiple search directions to be traced simultaneously. Numerical examples are presented to demonstrate the features of the proposed model and the computational efficiency of the developed algorithm.