本研究探討生產線平衡問題(Assembly Line Balancing Problem,ALBP),此問題的目的是在給定各工作的流程時間與工作站的週期時間(Cycle Time)下,如何有效地分配每個工作於工作站中,使總工作站數為最少。因此本研究探討的問題為平衡產線工作負荷,安排出最適當的工作調度於工作站,使其接近所設定之工作站週期時間,進而減少工作站數以降低企業成本,使產線的工作效率達到最大化。 本研究應用三種演算法,包含基因演算法(Genetic Algorithm, GA)、免疫演算法(Immune Algorithm, IA)以及粒子群演算法(Particle Swarm Optimization, PSO),並且提出全新的編碼方式解決此問題。測試問題分為兩部分,第一部分是以過去學者的部分資料庫做為測試問題,第二部分則是自行設計之測試問題。本研究將比較三種演算法對此生產線平衡中的工作站規劃問題的表現,數值結果顯示,粒子群演算法求解速度優於其他兩種演算法,而免疫演算法求解品質優於其他二種演算法。
This thesis explored the Assembly Line Balancing Problem. The purpose of this problem is to effectively allocate tasks to each workstation, subject to the given cycle time, such that the total number of workstations is minimized. Therefore, based upon the precedence of tasks, the present study aims to arrange the most appropriate tasks into workstations so close to the given cycle time of workstation and to reduce the number of workstations. A better assignment of tasks in Assembly Line Balancing Problem can reduce costs and incease the working efficiency. In this thesis, we apply three artificial intelligence algorithms, including Genetic Algorithm (GA), Immune Algorithm (IA), Particle Swarm Optimization (PSO), and a new encoding method for solving this Assembly Line Balancing Problem . There are two sets of test problems. The first set of test problems is adopted from the benchmark problems in the database, and the second set of test problems is our own designed set of test problems. In this thesis, we test these test problems and compare the performance of these three algorithms for assembly line balancing problems. Nnumerical results show that Particle Swarm Optimization is faster than the other two algorithms and Immune Algorithm is superior to the other two algorithms.