Using Grouping Genetic Algorithm to Solve Multi--Traveling Salesmen Problem

• Main Authors: Chien-chih Pai, 白健志
• Other Authors: Shih-hsin Chen
• Format: Others
• Language: zh-TW
• Published: 2011
• Online Access: http://ndltd.ncl.edu.tw/handle/71924458730653661548

碩士 === 南華大學 === 資訊管理學系碩士班 === 99 === 　　For most enterprises, they encounter the assignment problems for their own sales to visit different customers everyday. In addition, if the visiting sequence could be optimized, it is useful for enterprises to reduce their cost. This condition is actually the form of Multiple Traveling Salesmen Problem (mTSP), which is more complex than the Traveling Salesman Problem (TSP) . Among many approaches could solve the mTSP, this study attempts to use the grouping genetic algorithm (Grouping Genetic Algorithm, GGA) to solve this problem. Previous GGA is able to solve problem for the number of salesmen is not set in the beginning. However, due to the utilization concern, enterprises may fix the number of persons to be assigned. This thesis, thus, present two new mating operators and mutation operators value. The proposed GGA is further combined with a famous heuristic, Farthest Insertion, to reconstruct a sequence within a route. Extensive experiments were ran and then select better settings for the proposed GGA and compared it with the state-of-art algorithm. It could be interesting for the enterprises to employ the proposed algorithm due to its simplicity.