一种小种群自适应遗传算法研究

分析了变异算子在标准遗传算法和自适应遗传算法中的作用和当前研究的不足,提出一种新颖的能够大大提高遗传算法性能的变异策略,并进而提出一种小种群自适应遗传算法.该方法在采用赌轮选择和单点交叉的情况下,利用一种可伸缩的变异策略使得算法在探测和开发之间取得很好的平衡,从而能够用小规模的种群进行有效的全局搜索和局部搜索,避免早熟收敛,并能够以较快的速度收敛到全局最优解.对多峰函数的仿真实验表明了算法的有效性.

Research on Adaptive Genetic Algorithm with Small Population

The effect of mutation operator in simple genetic algorithm and adaptive genetic algorithm is analyzed,and the corresponding study is insufficient.A novel mutation strategy improving the performance of genetic algorithm greatly is introduced,and a new adaptive genetic algorithm with small population is proposed.The new algorithm adopted roulette wheel selection and one-point crossover makes the flexible mutation strategy obtain balance relatively between exploration and exploitation.The proposed method improves the global and local searching ability efficiently,avoids the premature convergence,and obtains the global optimal solution with a small population.The simulation about optimal problems of multimodal function shows the new algorithm is effective.