一种自适应紧凑遗传算法及其仿真研究

Adaptive Compact Genetic Algorithm and Simulation

首先提出了用联合熵来刻画紧凑遗传算法的多样性方法;在此基础上针对紧凑遗传算法存在的早期收敛,同时提出了基于多样性的自适应紧凑遗传算法。该算法通过种群多样性的变化和每个基因座自身的进化状态来控制概率向量的更新。这种更新策略不仅可以跟踪种群的全局进化状态,同时还可以对基因座自身的进化状态进行局部调整,从而提高了进化中种群的多样性和算法的搜索效率。通过典型函数的测试,仿真结果表明了提出的算法的优越性和有效性。

The method was proposed for measuring diversity of compact genetic algorithm by joint-entropy. An adaptive compact genetic algorithm based on population diversity was proposed for solving compact genetic excursion. This algorithm updates probability vector by variety of population diversity and evolutionary state of every gene. This updating mechanism not only tracks global evolutionary state of population but also locally adjusts evolutionary state of every gene. The algorithm improves population diversity and the searching efficiency. The experiments show advantages and efficiencies of this algorithm through testing with typical function.