摘要
在遗传算法中 ,群体多样性对于避免早熟收敛、提高搜索效率以及保证搜索的全局性是至关重要的。本文分析了遗传算法在进化处理过程中导致群体多样性下降的原因 ,并着重就交叉算子与替换策略 (或后期选择 )的共同作用对群体多样性的影响作了较深入的探讨 ,完成交叉算子关于群体多样性损失的量化分析 ,并据此对如何平衡算法的“探测”与“开发”性能做了讨论。
In genetic algorithms,population diversity is very important to avoid premature convergence and ensure the effectiveness and global property of search.This paper studies the reasons why the population deversity decreases gradually,and especially discusses the collective affection of crossover operators and replacement strategies.Some quantitative results about how crossover operation affects the population diversity are obtained.How to balance the performance between exploration and exploitation is discussed on the basis of the above results.
出处
《计算机工程与科学》
CSCD
2000年第4期46-49,共4页
Computer Engineering & Science
关键词
遗传算法
群体多性性
交叉算子
替换策略
genetic algorithms
population diversity
crossover operator
replacement strategy
