基于模式库的遗传算法在非静态函数优化中的应用

Nonstationary function optimization using genetic algorithms with a schema-base

本文针对非静态函数优化问题提出了一种基于模式库的改进遗传算法,此算法最主要的特点就是采用模式库来保存算法进行过程中的一些好的模式,这些模式一方面用来组建一些适合环境的优良个体,提高收敛速度,另一方面可以给早熟群体注入新的基因,增加群体的多样性。文中分别就函数优化问题和背包问题给出了改进遗传算法的仿真结果,表明本文算法的有效性。从理论上来说,本文算法适用于多个状态之间的动态变化环境。

A genetic algorithm with a schema-base for the non-stationary function optimization was proposed, in which the schema-base was used to save some good schemata in order to construct some optimal individuals, and then the convergence speed and the diversity of population were improved. At last the simulation tests on a nonstationary function optimization problem and a dynamic knapsack problem were made, which illustrated the validity of this algorithm in dynamic environrnent