An archived multi-objective simulated annealing for a dynamic cellular manufacturing system

Subject Areas : Mathematical Optimization

Hossein Shirazi ¹ , Reza Kia ² , Nikbakhsh Javadian ³ , Reza Tavakkoli-Moghaddam ⁴

1 - Department of Industrial Management, Qom Branch, Islamic Azad University, P.O. Box 3749113191, Qom, Iran
2 - Department of Industrial Engineering, Firoozkooh Branch, Islamic Azad University, P.O. Box 148, Firoozkooh, Iran
3 - Department of Industrial Engineering, Mazandaran University of Science and Technology, P.O. Box 734, Babol, Iran
4 - School of Industrial Engineering and Engineering Optimization Research Group, College of Engineering, University of Tehran, P.O.Box 11155-4563, Tehran, Iran

Received: 2020-10-02 Accepted : 2020-10-02 Published : 2014-06-01

Keywords: Simulated Annealing, Dynamic cellular manufacturing systems , Group layout  Production planning  Archived multiobjective,

Abstract :

To design a group layout of a cellular manufacturing system (CMS) in a dynamic environment, a multi-objective mixed-integer non-linear programming model is developed. The model integrates cell formation, group layout and production planning (PP) as three interrelated decisions involved in the design of a CMS. This paper provides an extensive coverage of important manufacturing features used in the design of CMSs and enhances the flexibility of an existing model in handling the fluctuations of part demands more economically by adding machine depot and PP decisions. Two conflicting objectives to be minimized are the total costs and the imbalance of workload among cells. As the considered objectives in this model are in conflict with each other, an archived multi-objective simulated annealing (AMOSA) algorithm is designed to find Pareto-optimal solutions. Matrix-based solution representation, a heuristic procedure generating an initial and feasible solution and efficient mutation operators are the advantages of the designed AMOSA. To demonstrate the efficiency of the proposed algorithm, the performance of AMOSA is compared with an exact algorithm (i.e., [-constraint method) solved by the GAMS software and a well-known evolutionary algorithm, namely NSGAII for some randomly generated problems based on some comparison metrics. The obtained results show that the designed AMOSA can obtain satisfactory solutions for the multi-objective model.