针对混合蛙跳算法SFLA(shuffled frog leaping algorithm)易陷入局部最优、收敛速度慢的问题,提出一种改进的混合蛙跳算法.该算法首先用混沌的Tent序列初始化青蛙群体以增强群体的多样性,提高初始解的质量；再根据每只青蛙的群体适应度方差值选取不同的变异概率,有效增强了SFLA跳出局部最优解的能力.通过对6个经典函数的仿真测试,结果表明,新算法比SFLA和ISFLA1的寻优能力更强,迭代次数更少,解的精度更高.%To address the problems of shuffled frog leaping algorithm (SFLA) such as prone to running into local optimal and slow in convergence, an improved shuffled frog leaping algorithm is presented. In this algorithm, first the frog population is initialised with chaotic Tent sequence to enhance the diversity of the group for improving the quality of initial solution. Then the different mutation probability will be selected according to population fitness variance of each frog, so that the ability of shuffled frog leaping algorithm is effectively enhanced in jumping out of local optimal solution. The simulation results of experiments on six classical functions show that compared with SFLA and ISF-LA1, the new algorithm has stronger searching ability, its iterations number is less, and the precision of solution is higher.
展开▼
