智能化遗传算法

针对遗传算法的收敛速度慢、收敛早熟和概率稳定性差等问题提出一种智能化遗传算法(IGA)。首先，建立描述种群进化的统计特征量，为IGA的算法策略提供决策依据。其次，建立种群的自学习算法、种群的自组织算法与遗传算子操作概率的自适应算法，并将这些算法嵌入最优保存简单遗传算法(OMSGA)，从而构成IGA。最后，从理论上对算法收敛性及效率进行了分析。通过遗传算法标准测试函数的仿真结果证明了算法的实用性和有效性。

An Intelligentizing Genetic Algorithm

The paper presents an intelligentizing genetic algorithm (IGA) to solve the problems of slow global convergence, quick prematurity and bad probability stability to genetic algorithms (GAs). First, the characteristic statistic variables of population evolution are established so that they are strategically used to the different steps of IGA. Afterwards, the self-learning algorithm, the self-organizing algorithm of population and the self-adapting algorithm of operational probability of operator are proposed, and they are embedded in optimum maintaining simple GA (OMSGA) to form IGA. Finally, the global convergence and computational efficiency are analyzed theoretically, and the paper demonstrates the practicability and validity of IGA by standard testing functions of GAs.