带记忆功能的混合蛙跳算法

Shuffled frog leaping algorithm with memory function

针对基本混合蛙跳算法收敛速度慢、优化精度低的问题、提出了带记忆功能的混合蛙跳算法。引入自适应学习因子,使算法在迭代初期加速收敛并不断拓展新的搜索区域,在迭代后期能够在全局最优邻域进行精细搜索,从而保持了开发与探索的平衡,并提高了收敛精度;采用随机分组策略平衡各子群的寻优能力,维持了种群的多样性。对6个测试函数进行了优化实验,并与基本混合蛙跳算法和相关文献中的改进算法进行比较,结果表明了该算法具有更好的优化性能。

Aiming at the problems of the shuffled frog leaping algorithm（SFLA）,such as slow convergence speed and low optimization precision,the shuffled frog leaping algorithm with memory function（MSFLA） is proposed.Through introducing the adaptive learning operator,make the algorithm convergence with faster speed and expand search area in early iterations,and search precisely in the global optimal neighborhood in later iterations,which kept the balance between exploration and development and improved the convergence precision.By using randomized grouping strategy,the optimization ability of memeplexes are balanced,and the population diversity is kept.Through testing six benchmark functions,basic SFLA and the improved SFLA are compared in related references,the results show that MSFLA had better performance.