【作者】 马红梅； 巩敦卫； 郭广颂；

【Author】 MA Hong-mei GONG Dun-wei GUO Guang-song School of Information and Electrical Engineering,China University of Mining & Technology,Xuzhou 221008,China.

【机构】 中国矿业大学 信息与电气工程学院；

【摘要】 针对现有的漂移平衡遗传算法的核种群规模较大,且其实现方式是串行的,设计了一种对等式并行漂移平衡遗传算法,该算法将核种群分成多个规模较小的核种群,以增强核种群的局部搜索能力,便于其在局部进行深度搜索,提高求解质量;然后给出了算法基于 Linux 集群的实现方案,计算了算法的执行时间及加速比,分析了影响算法性能的主要因素;最后通过复杂多蜂函数优化验证了算法的优越性．更多还原

【Abstract】 For the problems that the core population of the shifting balance genetic algorithm is large and its implement is sequential,a parallel shifting balance algorithm with peer-to-peer model is proposed to solve these questions.In this model,the core population is divided into several small core populations to enhance its capability of exploiting,which can exploit in local area deeply and obtain better solutions.Then a scheme of implementation on Linux cluster is presented,the speedup formula of the proposed algorithm is calculated,and the key factors that influence the performance of the parallel algorithm are analyzed.Finally,the algorithm’s efficiency is validated through optimization of a complicated multi-modal function.更多还原