Applying multi-agent mechanism for mixed-model assembly line balancing

• Main Authors: Jia-Hong Huang, 黃嘉宏
• Other Authors: Ching-Jen Huang
• Format: Others
• Language: zh-TW
• Published: 2013
• Online Access: http://ndltd.ncl.edu.tw/handle/19695507485703611040

碩士 === 國立勤益科技大學 === 工業工程與管理系 === 101 === In order to satisfy the different requirements of customers, the assembly line model have to change from producing single type production to producing multiple type one simultaneously on the same assembly line. And assigning jobs to stations becomes more complicate than before, this makes the assembly line balancing problem becomes an important issue. The idle time and overload of the labors or machines will be reduced if the assembly lines are balanced and then the production efficiency will be move up and the total synergy of assembly line will be improved. Differing from the heuristic or hybrid meta-heuristic used before, this study proposes a multi-agent system for mixed-model assembly line balancing. The system consists of planning agent, balancing agent and machine agent. The agents can negotiate between them iteratively to improve jobs assignment. The greatest positional weight is applied to obtain initial solution, and then the neighborhood search of Tabu search is employed to find the best move strategy to insert or swap jobs among machines. The results show that the proposed agent negotiation approach can minimize the work load variation and idle time of workstation on assembly line and achieve the goal of assembly line balancing.