基于t分布变异的蝙蝠算法

Bat algorithm based on t distribution mutation

为了进一步提高BA算法的性能,提出一种基于t分布变异的蝙蝠算法(TMBA).该算法通过对最优的蝙蝠个体进行高斯变异,对非最优蝙蝠个体进行自适应t分布变异,使得算法在进化初期具有良好的全局探索性,而在进化后期具有较优的局部开发性.通过选取6个典型函数对BA、ABA和TMBA进行对比实验,结果表明TMBA优于BA、ABA.

In order to improve the performance of algorithm,a Bat algorithm based on t distribu- tion mutation （TMBA） is presented. This improved algorithm executes the Gauss mutation on the excellent bat and executes the adaptivet distribution mutation on the nonexcellent bat. The proposed algorithm shows good exploitative properties at the early evolution and more explora- tive at later evolution process. It uses BA,ABA and TMBA to carry out numerical experiments for 6 test benchmarks. The simulation results show that the proposed TMBA is superior to BA and ABA.