摘要
为了充分发掘混合蛙跳算法求解复杂优化问题的能力,提出了一种新颖的改进混合蛙跳算法。改进算法借鉴粒子群优化算法的速度更新方式,通过族群中随机个体、最优个体和最差个体间的位置关系来确定最差个体的更新步长;借鉴差分进化思想,通过伪差分变异产生虚拟个体来更新最差个体,以提高种群开拓能力。通过对四个典型测试函数的仿真实验表明,相比其他几种改进算法,改进算法以100%的概率找到了某些函数的理论最优值,寻优效果更好,收敛成功率更高。
In order to fully exploit the ability of shuffled frog leaping algorithm (SFLA) to solve complex optimization prob- lems, this paper proposed a novel improved shuffled frog leaping algorithm. It adopted the evolutionary methods of particle swarm optimization(PSO) algorithm,and used the relative position among the random individuals, the best individual and the worst individual in subpopulation as the first update step. And in order to improve the population extension capacity, it con- structed a virtual individual by pseudo-differential mutation to replace the worst individual of subpopulation. The simulation re- suhs of four typical test functions show that, comparing to the other improved algorithms, the proposed algorithm can find the theoretical value of certain functions by probability of 100%, and has better optimization results and better success rate.
出处
《计算机应用研究》
CSCD
北大核心
2014年第9期2681-2684,共4页
Application Research of Computers
基金
航空科学基金资助项目(2013ZC53038)
关键词
函数优化
混合蛙跳算法
粒子群优化
速度更新
伪差分变异
function optimization
shuffled frog leaping algorithm(SFLA)
particle swarm optimization( PSO )
velocity up-date
pseudo-differential mutation