Development of a comprehensive model and BFO algorithm for a dynamic cellular manufacturing system

• Choosing the machines with higher capability and lower relatively cost.
• Less backtracking and material handling, and work-in-process (WIP) inventory.
• Increasing shop flexibility and machine utilization.
• Optimizing doing tasks for completing operations of CMS.
• Modelling and simulating the real production environments.

A comprehensive multiobjective model of the cellular manufacturing system (CMS) operating in a dynamic environment is developed. The proposed algorithm takes into consideration various important cell design issues, such as machine assignment, intercell/intracell material handling, worker assignment, outsourcing and workload balancing based on operational time and operation sequence of the parts. Workload balancing among cells tends to increase CMS processing costs, and hence the use of multiobjective optimization methods is required for CMS solutions that are optimal and feasible. A multiobjective matrix-based bacteria foraging optimization algorithm with traced constraint handling (MOMBATCH) is developed for this purpose. The performance of the proposed algorithm is compared with that of the non-dominated sorting genetic algorithm II method that is frequently reported in the literature and the off-the-shelf program CPLEX. The results show that MOMBATCH solves problems more efficiently in terms of finding optimal solutions while maintaining the Pareto frontier diversity.