摘要
面对经典蝙蝠算法(BAT ALGORITHM,BAT)容易过早收敛、收敛效果差的问题,提出一种带有期望因子的新型蝙蝠算法,将蝙蝠的觅食行和L′evy飞行特征相结合来模拟蝙蝠的捕食过程,并在迭代公式中加入期望因子,进而采用全新的更新频率、速度、位置的方式,再利用L′evy的飞行特性使得该新型算法更容易趋向目标最优值。通过标准测试函数对其提出的算法进行仿真测试,并与布谷鸟搜索算法、粒子群算法、基本蝙蝠算法相比较。仿真结果表明该算法增强了原算法的收敛精度以及寻找最优目标的能力,性能明显优于其他3种算法,是一种优化高度非线性、复杂函数问题的有效手段。
Facing the shortcoming of the classic bat algorithm which is easy to premature convergence and the convergence precision is low. We put forward a new bat algorithm, combining bats' foraging behavior with L′evy flying characteristics to simulate predation process of bats, adding the expectation factors into the iteration formula, adopting the new way to update frequency, speed and location and using L′evy flight characteristics to make the new algorithm easier to tend to the optimal target value. The proposed algorithm is tested through standard test functions. The simulation results show that the algorithm improves the astringency and ability to search for the optimal target of the original algorithm. The performance is better than the other three algorithms. It is a good method of solving highly nonlinear and complex function optimization problems.
出处
《控制工程》
CSCD
北大核心
2016年第S1期83-87,共5页
Control Engineering of China
