Applying multi-agent approach to mixed-model assembly line balancing

Abstract

This paper develops a decentralized approach using multi-agent based framework to solve the mixed-model assembly line problem. A three-level multi-agent system is proposed. Tabu search technique is applied in the line balancing process. A modified ranked positional weight is developed and used to produce the initial solution. And a restricted neighborhood strategy is proposed to flexibly adjust the workloads of all machines. Two criteria are simultaneously considered for optimization: to minimize the number of workstation for a given cycle time, and to maximize the workload smoothing.