一种基于遗传算法的进化计算模型

提出一种基于遗传算法的进化计算模型(ECM).在ECM的种群中,每个成员都根据其适应度值不同程度地影响着种群的进化.ECM定义了个体对进化的影响因子,并以个体的影响因子为参数定义了个体的形成算子.分析表明,ECM是采用算术交叉算子的两父辈遗传算法以及采用频率扫描交叉算子的多父辈交叉遗传算法的推广,形成操作是父代群体编码的凸组合.实验研究显示,ECM具有比经典遗传算法更强的优化计算功能.

Evolutionary computing model based on genetic algorithm

An evolutionary computing model,ensemble crossover model (ECM) based on genetic algorithm (GA), is introduced. In ECM, every individual in the parental population impacts the evolution to a certain extent. Every individual is given an evolutionary impact factor according to its fitness. The offspring forming operator is defined on the impact factors of parental population. Analysis shows that ECM is the generalization of both the two-parent GA with arithmetic crossover operator and the multi-parent recombination GA with occurrence based scanning crossover operator. And the forming operation is the convex combination of the whole parental population. Experiments show that ECM outperforms the classical two-parent GA.