自适应变异的蝙蝠算法

针对基本蝙蝠算法(BA)寻优精度不高、收敛速度慢和易早熟收敛的问题,提出一种改进的具有自适应变异机制的蝙蝠算法,用以求解复杂函数问题。利用K-means聚类对蝙蝠种群进行初始化,使种群在搜索空间分布更为均匀；采用根据迭代次数自适应变化的控制概率Pt判断算法是否进行高斯变异,增强种群多样性,促使蝙蝠个体跳出局部极值点；将自然选择思想引入BA,提高算法搜索速度,避免早熟收敛。选取几个典型函数进行测试,结果表明改进算法优化性能有了显著提高,具有较快的收敛速度,较高的寻优精度、收敛稳定性和收敛可靠性,验证了改进蝙蝠算法(IBA)的有效性及优越性。

Bat algorithm with adaptive mutation

In order to overcome the problems of low convergence precision,convergence speed and easily relapsing into the premature covergence in basic Bat Algorithm(BA), an Improved Bat Algorithm(IBA) with adaptive mutation mechanism is proposed to solve complex function optimization problems.Firstly, initializing the bat population by K-means clustering, to make the poulation distributes more umiform. Sencondly, in order to escape from lacal optima, introduceing the Gaussian mutation, whether the mutation is happened controlled by the adaptive mutation probability Pt which is changed with the number of interation. And introduce the idea of natural selection into the bat algorithm to improve the algorithm search speed and overcome the shortcoming of BA that it is easy to fall into the premature covergence.Experimental results show that the IBA improved the global optimizing ability remarkably which includes optimizing accuracy,convergence speed,stability and reliability .