一种求解TSP问题的改进遗传算法杨华芬.pdfVIP

第21 卷 第5期 重 庆 工 学 院 学 报( 自然科学版) 2007 年5 月 Vol.21 No.5 Journal of Chongqing Institute of Technology(Natural Science Edition) May 2007 TSP 1,2 1 杨华芬 ,魏 延 ( 1. , 400047;2., 655000) : (Traveling Sales an Proble TSP) , TSP . TSP , : , 2_opt ,, , . , , . 20TSP , . :TSP; ;2_opt ;; :TP301. 6 :A : 1671- 0924( 2007)05- 0086- 05 Improved Genetic Algorithm for TSP 1, 2 1 YANG Hua_fen , WEI Yan ( 1. School of Mathe atics and Co puter Science, Chongqing Nor al University, Chongqing 400047, China; 2. Qujing Nor al College, Qujing 655000, China) Abstract: Traveling Sales an Proble (TSP) is a classical co binatorial opti ization proble . Genetic algorith (GA) is a ethod for solving this proble . But it is hard for GA to find global opti ization quickly and prevent pre ature convergence. Consideringthe characteristic of TSP, this paper puts forward an i proved genetic algorith , that is, using greedy strategy to initialize species, using 2_opt operator to opti ize it so that the initialized individuals contain opti ized paths, which quickens the convergence of the algorith to so edegree and protects pre aturity and closerelative reproduction. After i proving cross operator and utation operator, the diversity of population and ost of thegood perfor anceof last genera tion can be aintained. And the i proved algorith is used tosolveTSP proble in 20 cities and the ex peri ental result indicates that the algorith is fast and has good solutions. Key words: TSP; cro