基于二次插值的天牛须搜索算法

Beetle antennae search based on quadratic interpolation

针对天牛须搜索算法在高维空间中搜索精度低和易陷入局部最优的问题进行了研究,提出一种新的天牛须优化算法——基于二次插值的天牛须搜索算法(QIBAS)。算法在天牛进行移动后,将天牛当前位置左右两触须作为插值坐标点,利用二次插值生成一个新的解,再对比插值产生的解与当前最优解、全局最优解的适应度值,更新全局最优解。对多个单峰函数和多峰函数进行数值仿真测试,其维度分别取100、500、1000、5000、10000。仿真结果表明,引入二次插值有效提升了BAS算法跳出局部最优的能力。QIBAS在求解最优值时,其求解精度有极大的提升,收敛速度也有较明显提升,改进算法的有效性得以验证。

Focused on the problems that low-precision and easily to fall into local optimum result in the high-dimensional space by the longhorn search algorithm,this paper proposed a new optimization algorithm of the short-lived beetles-the short-lived search algorithm based on quadratic interpolation which was called QIBAS.The algorithm took tentacle’s left and right tentacles as interpolation coordinate points after the beetle moved.And it generated new solution with a second interpolation,then compared to the current optimal solution and the global optimal solution with the fitness of interpolation solution.At the same time to update the global optimal solution,it performed numerical simulation tests on multiple unimodal and multimodal functions,and their dimensions took as 100,500,1000,5000 and 10000.The simulation results show that the introduction of quadratic interpolation effectively improves the ability of the BAS algorithm to jump out of the local optimum.When QIBAS solves the optimal value,its solution accuracy is greatly improved,and the convergence speed is also significantly improved.The effectiveness of the improved algorithm is verified.