并行遗传算法收敛性分析及优化

针对并行遗传算法（parallel genetic algorithms，PGA）容易出现收敛过快和陷入局部最优解的问题，综合多种不同进化策略遗传算法之所长，设计了一种混合的粗粒度并行遗传算法。该算法由多个独立的子群体组成，各个子群体并行的、独立的、按照不同的遗传进化策略进化，每隔一定的时间，在子群体之间进行最优个体的迁移，促进群体的共同进化，并抑制群体早熟。在PVM环境下，用该算法实现函数优化问题，仿真实验数据表明了其有效性．

Convergence analysis and optimization of Parallel Genetic Algorithm

Aim at the problems of which Parallel Genetic Algorithms is very easy convergence and getting in the local best individual, colligate the strongpoint of manifold genetic algorithms which ttas different evolutionary strategy, a novel Parallel Hybrid Genetic Algorithm is proposed in this paper. It is compose of multi-population. Diferent population evolves independently with different genetic evolutionary strategy. It migrates the best individual between different population termly to accelerate all population together evolving and restrain the population premature phenomena. Under PVM environment, it has been proved that the algorithm is effective by the experiment of settling the question of function optimize.