一种小生境正交遗传算法研究

Research on a niche orthogonal genetic algorithm

针对标准遗传算法的不足 ,借助正交试验法的全局均衡设计思想和二元变异操作对初始种群产生方式、交叉算子和变异算子进行了改进 ,提高了种群的多样性 ;借助最优保留策略和自然界的小生境思想 ,对选择算子进行了改进 ,提高了算法的全局收敛性能 ;另外还通过引入加速正交搜索操作 ,提高了算法的收敛速度 .在此基础上 ,提出了一种小生境正交遗传算法 ,并进行了实例研究 .研究结果表明 ,该算法不但可以有效地克服标准遗传算法的缺陷 ,而且计算速度。

Targeting the insufficiencies of the standard genetic algorithm (SGA), improvements are made to SGA from three aspects in this paper. Firstly, for the purpose of enhancing diversity of population, production mode of initial population, crossover operator and mutation operator are improved by using the global equilibrium design ideology of orthogonal experiment method and dual mutation operation; then for the purpose of enhancing global convergence performance of algorithm, the selection operator is improved by using elitist preservation strategy and natural niche ideology; finally, for the purpose of enhancing convergence speed of the algorithm, an accelerated orthogonal search operation is imported. On this basis, a niche orthogonal genetic algorithm (NOGA) is proposed in this paper, and case research is also done. The research results show that NOGA can not only overcome the shortcomings of SGA effectively, but also evidently improve the computing speed, computing precision and stability of the algorithm.