一种新的并行遗传算法应用研究  

Research of an Improved Parallel Genetic Algorithm with Adjustable Migration Rate

作　　者：梁旭[1] 赵戈[1] 黄明[1] 

机构地区：[1]大连交通大学软件学院,辽宁大连116028

出　　处：《化工自动化及仪表》2009年第1期31-34,共4页Control and Instruments in Chemical Industry

摘　　要：针对传统粗粒度并行遗传算法容易早熟并且收敛速度较慢的不足,提出一种新的并行遗传算法。新算法采用了一种基于相对平均适应度的动态调整迁移率的方法及时增加了其它岛屿到相对平均适应度较高的岛屿的迁移率,从而有效的增加了种群的多样性,抑制了早熟现象,提高了最优解的质量。同时,提出了一种存活期的计算方法,对直接迁入每个岛屿中的个体计算存活期并淘汰年龄超过存活期的个体,控制了岛屿的规模,增强了算法的收敛性能,加快了收敛速度。将提出的改进算法用Muth and Thompson基准问题测试,验证了该算法的有效性。Aiming at the insufficiency of traditional coarse grain parallel genetic algorithm both in gence and long execution time, a new genetic algorithm with adjustable migration rate （AMRPGA） one hand, it dynamically increases the comparative migration rate of other islands to the island with premature conver- higher average fit- ness and reduces the possibility of premature convergence and improves the quality of the final result by introducing a way of adjusting migration rate dynamically. On the other hand, it controls the size of the population and improves the convergence capability of the algorithm and increase the convergence speed as well, through calculating the lifespan of those individuals directly migrated to the island. This improved genetic algorithm has been tested by Muth and Thompson basic problem and its validity is revealed.

关 键 词：迁移率 粗粒度 存活期 相对平均适应度 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]