Hybrid Simulated Annealing and Particle Swarm Optimization Algorithms for Traveling Salesman Problems

• Main Authors: Kuo-Lung Liao, 廖國隆
• Other Authors: Hsuan-Ming Feng
• Format: Others
• Language: zh-TW
• Published: 2008
• Online Access: http://ndltd.ncl.edu.tw/handle/10377006719037304935

碩士 === 國立金門技術學院 === 電資研究所 === 97 === Traveling Salesman Problem (TSP) like the typical discrete scheduling, assigning system can be considered as an NP-Complete search problem. The heuristic Particle Swarm Optimization (PSO) algorithm is first introduced by Kennedy and Eberhart in 1995. Based on the global search capability of PSO, it has been proven to well solve continuous optimal problems. In this literature, the applied novel PSO-TS-SA algorithm with PSO, Transfer Space (TS) and Simulated Annealing (SA) schemes is integrated to find near optimal solutions for several type cities TSP problems. The Transfer Space (TS) operation in such sorting way is proposed to recover the continuous results. Simulated Annealing (SA) aimed at avoiding the trapped in local optimal condition is applied to improve the jump ability of the flying particle. Moreover, a popular Fuzzy C-means Clustering (FCM) algorithm is used to divide large TSP cities into suitable traveling groups. A novel merging algorithm is acted as a connection machine to connect the head and end of the selected cities groups. Its objective is to recovery the near optimal traveling path in a shorter spending time. Several TSP testing data set is used to demonstrate the adaptation of the proposed PSO-TS-SA in some experiment results, the PSO-TS-SA algorithm compared with other learning algorithm is illustrated its better performance. In other experiments, the popular FCM algorithm the proposed FCM clustering and merging based learning algorithm can fast approximate the desired output in large traveling salesman problem.