一种求解函数优化的混合差分演化算法

Hybrid Differential Evolutionary Algorithm for Solving Function Optimization Problems

为了解决传统遗传算法易陷入局部最优解的问题,在多父体杂交算法和差分进化算法的基础上,提出了混合差分演化算法。该算法的核心在于,采用多父体杂交算子保证算法的遍历性,通过淘汰相同个体来保持群体的多样性,并以较小概率随机选取部分个体进行差分进化操作,从而充分利用最优个体的信息达到了加快收敛速度的目的。对复杂函数的寻优实验验证了混合差分演化算法的有效性。

A hybrid differential evolutionary algorithm was proposed to avoid trapping local optimum. The algorithm is based on multi-parent crossover and differential evolution, and the key points of it lie in： 1） use multi-parent crossover to ensure ergodicity; 2） remove identical individuals from the population for maintaining the diversity; 3） select individuals with low probability to evolve using differential evolution operator, as a result of this, the information of the best individual can be used to speed up the evolution. Experimental results on the complex function show that this algorithm is efficient.