融合改进天牛须和正余弦的双重搜索优化算法

Dual Search Optimization Algorithm Based on Improved Beetle Antennae Search and Sine Cosine Algorithm

针对标准正余弦优化算法在搜索时存在的计算精度低、容易陷入局部最优值等缺点,借鉴天牛须算法运算简单、搜索速度较快的特点,本文提出了一种融合改进天牛须和正余弦的双重搜索优化算法(BAS-SCA).首先,在现有标准正余弦算法基础上,引入动态自适应权重机制来平衡全局搜索和局部搜索,提升收敛速度;其次,提出了一种新的转换参数模型,该参数模型通过结合指数型函数和余弦函数来替换传统的线性衰减函数;最后,为了提升正余弦的搜索精度和速度,同时尽可能跳出局部最优解,提出了改进的天牛须搜索算法,引入动态步长搜索机制将固定步长搜索改为变步长搜索,创新性地将改进的天牛须算法与改进的正余弦算法进行融合实现双重搜索优化,有效避免局部极值问题.实验表明,通过14个标准测试函数验证,所提双重搜索优化算法BAS-SCA相较于其它现有优化算法,具有更高的寻优精度和更快的收敛速度.

Aiming at the shortcomings of standard sine and cosine optimization algorithm,such as low calculation accuracy and easy to fall into local optimum value,in this paper,based on the advantages of beetle antennae search algorithm in terms of simple operation and fast speed in search,a dual search optimization algorithm(BAS-SCA)is proposed by integrating the improved sine and cosine optimization algorithm and improved beetle antennae search algorithm.First,dynamic adaptive weight mechanism is adopted to balance global search and local search,as well as improving the convergence speed;Second,a new conversion parameter model is proposed to replace the traditional linear decay function by combining exponential function and cosine function.Finally,in order to obtain better solution in a short time and avoid the local extreme problem,fixed search step in the traditional beetle antennae search algorithm is replaced by designed variable step strategy.The improved beetle antennae search algorithm and the improved sine cosine algorithm are combined to achieve double search optimization innovatively.Experimental results on fourteen standard test functions indicate that the proposed dual search algorithm has higher search accuracy and faster convergence rate comparing with other existing optimization algorithms.