摘要
设计了基于近邻点初始化和改进Inver-over(反序杂交)算子求解旅行商问题的并行演化算法。该算法执行时,主进程每当收集到各个种群的最好个体并形成精英种群时,就对该种群执行一次Inver-over算子,然后将其中最好的个体发送给各个种群。在PVM(并行虚拟机)并行环境下的实验结果表明,并行后能取得更好的解,并且在主进程中建立精英种群的演化有助于更好更快的收敛。
A parallel evolutionary algorithm for TSP,which is based on nearest neighbor initialization and improved Inver-over operator,is purposed.In this algorithm,once the master process has received all the best individuals from each population,it will generate an elite population and run Inver-over operator once.Then,send the best one to each sub-population.The experimental result based on PVM(parallel virtual machine)shows that the parallel algorithm get more reasonable solution and the elite population contributes to the convergence of the evolution.
作者
谢大同
李程俊
康立山
XIE Da-tong;LI Cheng-jun;KANG Li-shan(School of Computer,China University of Geosciences,Wuhan.430074,China;State Key Laboratory of Software Engineering,Wuhan University,Wuhan 430072,China)
出处
《计算机工程与设计》
CSCD
北大核心
2007年第10期2248-2249,2279,共3页
Computer Engineering and Design
基金
国家自然科学基金项目(60473081)
中国地质大学(武汉)2006年度优秀青年教师资助计划基金项目(CUGQNL0633)
关键词
旅行商问题
反序杂交
并行
主从式
近邻
精英种群
TSP
inver-over
parallel
master-slave
nearest neighbor
elite population
