一种用于智能组卷的自适应小生境遗传算法

在采用遗传算法进行智能组卷的过程中,常出现选择策略缺乏多样性保护机制的现象,易出现早熟收敛。为解决智能组卷的早熟收敛问题,提出一种自适应小生境遗传算法。采用小生境技术可提高个体的选择概率,增加个体的多样性选择机率;在保证算法收敛速度的情况下,给出一种一致变异算子,同时调节个体的变异概率和变异范围,提高种群多样性。最后,通过具体实例验证了该算法在较短的组卷时间内,可以实现全局寻优的结果,从而证明该算法的有效性。自适应小生境遗传算法在智能组卷中的应用具有实际意义。

A Kind of Adaptive Niche Genetic Algorithm for Intelligent Test Generation

In the process of the genetic algorithm for the intelligent test generation,it often lacks of diversity protection mechanism about selection strategy,it easily appears premature convergence.To solve the problem of intelligent generation,this paper proposes an adaptive niche genetic algorithm.This algorithm uses a niche technology to optimize the selection probability in order to improve the using efficiency of individuals.To improve the diversity of population without compromising the convergence speed,a uniform mutation operator also is designed by adjusting the mutation probability and the mutation range to improve the population diversity.Finally,Through concrete examples,the algorithm is verified in a short test time to realize the global optimization results,thus it proves the validity of the algorithm.The adaptive niche genetic algorithm application in test paper generation has practical significance.