AN ADAPTIVE MULTI-AGENT SYSTEM FOR STOCHASTIC ASSEMBLY LINE BALANCING
In this paper, we propose a multi-agent system having adaptable ability for assembly line balancing prob lems with stochastic task processinlgg time, and the proposed system can solve sinlggle-model assembly line balancing problems usinlgg information from adjacent agents. In the proposed system, workers are regarded as agents, and are assumed to know precedence relations of tasks to produce the product, and two adjacent agents exchange information about the allocated tasks and the total task time in sinlggle flow line. To balance the workload, an agent having less workload sends a request message to its neighbor agent having much more work load to exchange tasks between them. There is previous research usinlgg multi-agent system for stochastic assembly line balancing 1, However, it was required large number of iterations in the simulation experiment because there was not any mechanism to stop exchange tasks between agents. Therefore, agents continue to exchange tasks until the number of iterations fixed by parameters. To solve this problem, we add some mechanisms that judge exchange stop from own production environment, etc. Based on the background, this research develops an adaptive multi-agent system based on information of adjacent agents to reduce the number of iterations. Simulation experiments show the performance of the proposed system by comparing the accuracy of solutions and the number of iterations with other centralized methods and previous method.
Multi-Agent System Assembly Line Balancing Stochastic Decentralized Approach
Kana Yokoyama Katsumi Morikawa Katsuhiko Takahashi
Department of Artificial Complex Systems Engineering,Graduate School of Engineering, Hiroshima University, 1-4-1,Kagamiyama, Higashi-Hiroshima 739 - 8527, Japan
国际会议
The Ninth International Conference on Industrial Management(第九届工业管理国际会议 ICIM2008)
日本大阪
英文
120-129
2008-09-16（万方平台首次上网日期，不代表论文的发表时间）