• 中国计算机学会会刊
  • 中国科技核心期刊
  • 中文核心期刊

J4 ›› 2008, Vol. 30 ›› Issue (4): 73-75.

• 论文 • 上一篇    下一篇

群体间竞争与协作的遗传算法解决旅行商问题

庞留勇[1,2] 曹炬[1] 郑强[1]   

  • 出版日期:2008-04-01 发布日期:2010-05-19

  • Online:2008-04-01 Published:2010-05-19

摘要:

旅行商问题是经典的NP难组合优化问题之一,快速有效地解决旅行商问题具有重要的理论和实际意义。受自然界物种群体间相互联系的启发,提出了群体间竞争与协作的遗传算法来解决旅行商问题。该算法在迭代的过程中,每次只选择竞争力大的种群进行进化,同时为了维持各个种群间发展的平衡,对它们进行周期性的交流,能促使进化过程中中好的基因模式迅速地在各个种群中传播,提高了整体的进化速度。此算法不但能有效地维持群体的多样性,而且能提高收敛的速度。通过对旅行商问题的仿真实验,证明了该算法的可行性与有效性。

关键词: 旅行商问题 遗传算法 群体间竞争与协作 免疫选择

Abstract:

The classic travding salesman problem is one of the NP-hard combinatorial optimization problems, and a fast and effective solution to TSP has the impo  rtant theoretical and practical significance. Enlighted by the mutual contacts between the nature species groups, the paper proposes the inter-group com petition and collaboration of genetic algorithms to solve the traveling salesman problem. In the iterative process, we only choose a competitive group to execute the evolutionary process each time. Meanwhile in order to maintain the balance between all the groups, the populations have a cyclical exchang e, so as to promote good gene patterns to come into being in the process of olution. This algorithm is not only effective for maintaining the diversity of the population, and will increase the speed of convergence. Simulation results show that the algorithm is feasible and effective.

Key words: TSP;genetic algorithm;intergroup competition and collaboration;immune choice