用于多峰函数优化的改进跳跃基因遗传算法

跳跃基因是维持生物大脑神经细胞多样性的主要原因,因此在遗传算法中引入跳跃基因操作能够提高算法的全局搜索能力。然而,标准跳跃基因遗传算法的随机跳跃过程容易破坏较优性能染色体的基因。针对此问题,提出了一种改进跳跃基因遗传算法。在改进方案中,适应度越高的染色体上的跳跃基因,能以越高的概率朝性能比它差的染色体上跳跃,以提高进化速度。并且,在适应度函数中引入密度函数,以保持染色体的多样性。通过对经典多极值测试函数的寻优仿真表明,改进跳跃基因遗传算法能够更有效地提高遗传算法对复杂多峰函数最优解的求解速度与精度。

Improved Jumping Gene Genetic Algorithm for Multi-peak Function Optimization

The diversities of the brain neurons can be explained by jumping genes,thus the gene jumping operation is introduced to the traditional genetic algorithm.The global searching ability of the algorithm is improved.However,the random jumping operation of the standard jumping gene genetic algorithm(JGGA) destroyes the better genes easily.Therefore,this paper presents a modified JGGA.In the modification,the jumping genes on the chromosomes have higher fitness jump to the random selected chromosomes and worse fitness by higher probabilities to improve the evolution velocity.Furthermore,a density function is introduced to the fitness function to keep the diversities of the chromosomes.Finally,simulation results of the optimization of some multi-peak functions show that the modified JGGA can more effectively improve the speed and the precision for searching the optimal solution.