基于混合策略机制的人工鱼群算法

Artificial Fish-school Algorithm Based on Hybrid Strategy Mechanism

针对人工鱼群算法在处理多峰函数问题时存在一部分人工鱼处于漫无目的的随机移动、易陷入"早熟收敛"情况造成的收敛速度减慢的缺点,提出一种基于混合策略机制的人工鱼群算法。它借鉴群体位置方差的早熟判断机制,把云发生器产生的杂交和变异算子引入到该算法中,为减少算法计算量,而采用耗散的人工鱼群算法结构。仿真实验表明,该算法比只有一个适应值的人工鱼群算法具有更快的收敛速度,且具有很强的避免局部极小能力,其性能远远优于单一优化方法。

Aiming at the shortcoming of slow convergence rate caused by parts of artificial fish in the presence of randomly aimlessly move and tends to plunge into ＂premature convergence＂ situation when artificial fish-school algorithm dealt with multi-modal function,an artificial fish-school algorithm based on hybrid strategy mechanism was put forward.Referring to premature judgment mechanism of group position variance,the hybrid and variation operator were introduced into the algorithm.The simulation experiment showed that the algorithm has faster convergence speed than the artificial fish-school algorithm only with one fitness value,and has a strong ability to avoid local minimum,its performance is far superior to the single optimization method.