一种基于选择策略的差分混合蛙跳算法

摘要

设计了一种选择差分混合蛙跳算法SDSFLA,该算法通过增加组内个体更新个数提高了种群更新效率;通过引入差分进化算法的交叉算子和变异算子,加强了个体之间的信息交流;使用多种更新策略,提高了实验个体产生的成功率;随机选择控制参数,增加了种群的多样性.基于16个基准测试函数,将SDSFLA与一种改进的蛙跳算法、两种改进的差分进化算法进行对比,实验结果证实了SDSFLA算法的有效性和稳定性.%A Selected Differential Shuffled Frog Leaping Algorithm (SDSFLA) is proposed.Compared with the classic Shuffled Frog Leaping Algorithm (SFLA),SDSFLA improves the efficiency of populationupdatingby increasingthe number of updated vectors,uses thecrossover operator and mutation operator of Differential Evolution(DE) to strengthen information exchanges between vectors,uses multiple updating strategies to improve the success rate of the trial vectors,and increases the diversity of the population viarandomly-selected control parameters.Based on 16 benchmark functions,SDSFLA is compared with an improved frog leaping algorithm and two improved differential evolution algorithms.The test results confirm the validity and stability of the SDSFLA algorithm.