摘要
遗传算法是根据生物进化思想而启发得出的一种全局优化算法。通过求解TSP问题,对遗传算法的内部机理进行细致分析,给出一种基于主从式控制网络的并行遗传算法,同时对其内部遗传算子进行改进。通过各种遗传算子的优化组合,有效地控制了种群的早熟,并行计算实行异步通讯,时间复杂度上有明显改进。实验证明该算法具有很强的实效性,并具有良好的全局收敛性能。
GA (genetic algorithm) is a global optimizing algorithm inspired by the thought of biological evolution. This paper analyzes the inner mechanism by solving traveling salesman problem, and proposes a parallel algorithm based on master - slave control network. At the same time, we improve the inner GA operator. By organizing all kinds of operators properly, it can control premature effectively. The results of experiments show that this algorithm is practical and efficient.
出处
《计算机与数字工程》
2006年第11期1-4,8,共5页
Computer & Digital Engineering
基金
国家自然科学基金资助课题(编号:30370356
60373089)
关键词
TSP问题
遗传算法
并行计算
主从网络
TSP problem, genetic algorithm, parallel compute, master- slave network