Using Miners Genetic Algorithm to Solve Travelling Salesman Problems

• Main Authors: Bo-jhih Chen, 陳柏志
• Other Authors: none
• Format: Others
• Language: zh-TW
• Published: 2012
• Online Access: http://ndltd.ncl.edu.tw/handle/11066710453199853202

碩士 === 國立臺灣科技大學 === 工業管理系 === 100 === This thesis modifies the structure of Genetic Algorithms (GA) to develop a new algorithm to solve the Traveling Salesman Problem (TSP). The new algorithm is called Miners Genetic Algorithms (MGA). The MGA includes three phases as follows: (i) Mining, (ii) Exploration and (iii) Communication. The first phase (Mining) takes down all of the solutions on a mining location. The second phase (Exploration) searches new mining locations. The last phase (Communication) decides a mining location through the information of different mining. Since the MGA is a improve version of the GA, the MGA has the advantage of the GA. In addition, the MGA does not set the mutation and mating rates. In order to verify the efficiency of the MGA, the TSP examples are given to illustrate the efficiency of the MGA. Further, the efficiency of the MGA is compared with the GA for the TSP. In the TSP examples, we find that the initial value of MGA is produced using the Greed Algorithm and have more Explorations, which can find the better solution for TSP. Therefore, this method is called MGA_GM. In this thesis, the MGA_GM is compared with some heuristic algorithms for TSP. The comparison results show that the efficiency of MGA_GM is better than the other algorithms.