The most recent revolution in industry (Industrial Revolution 4.0) requires increased flexibility, agility and efficiency in the use of production equipment. The Dynamic Cellular Manufacturing System (DCMS) is one of the best production systems to meet such requirements. In addition, the increasing importance of environmental and social issues, along with recent laws, is forcing manufacturers and managers to take account of sustainability when designing and configuring manufacturing systems. This paper proposes a new bi-objective mathematical model of the Dynamic Cell Formation Problem (DCFP), in which both the worker's assignment and environmental and social criteria are considered. The first objective in this model is to minimize both production and labor costs, and total waste (e.g., energy, chemical material, raw material, CO2 emissions, etc.). Social criteria are represented as constraints. Due to the NP-hardness of this problem, we propose a new resolution approach called NSGA II-MOSA, that merges an efficient hybrid meta-heuristic based on the Non-dominated Sorting Genetic Algorithm (NSGA-II), with Multi-Objective Simulated Annealing (MOSA). Randomlygenerated test problems demonstrate the performance of our algorithm.
Literature reviewThis section gives an overview of the most prominent research on DCMS. Due to the large number of investigations in this area, we focus mainly on recent studies. First, Rheault et al. [3] introduced the concept of a dynamic environment in CFP. Schaller et al. [4] integrated CFP with inventory aspects, then showed the performance of their model on multiple heuristics and evaluated several alternative lower bounding methods. Chen and Cao [5] proposed a mathematical model for multi-period Cellular Manufacturing Systems (Dynamic CMS) minimizing the total cost, which includes: inter-cell material handling, inventory holding and the setting up of cells. They also developed a Tabu Search (TS) method to obtain good solutions and show the efficiency of their model. Next, these authors [6] generated a robust system configuration by integrating cell formation and part allocation. They also proposed a twostage TS to find the optimal or near optimal solutions. Tavakkoli-Moghaddam et al. [7] presented a nonlinear integer model of DCMS with machine capacity limitation, machine replication, inter-cell movements and production in batches. They used constant and variable costs as well as reconfiguration and inter-cell movement costs to formulate their objective function. Some of these authors [8] applied a Memetic Algorithm (MA) to solve their DCMS model. Defersha and Chen [9] formulated a comprehensive model containing dynamic cell configuration, alternative routings, lot splitting, sequence of operations and workload balancing. They also considered machine adjacency and cell size capacity as constraints. Moreover, Defersha and Chen [10] also proposed a two-phase GA-based heuristic to solve DCFP with alternative routings. Safaei et al. [11],[12] presented a DCMS mathematical model with unc...
