求解高维复杂函数的两类新型蝗虫优化算法

Two Novel Types of Grasshopper Optimization Algorithms for Solving High-Dimensional Complex Functions

针对蝗虫优化算法(GOA)在求解高维复杂函数的全局最优值时收敛精度不高、收敛速度较慢的问题,给出两类基于非线性权重的新型蝗虫优化算法(NGOA)。NGOA1主要改进了位置更新公式中的线性自适应参数,并将其作为当前最优位置的权重,以利于平衡算法的探索与开发能力。NGOA2是在NGOA1的基础上,给出精英蝗虫位置更新策略。该方法提高了算法的收敛速度,且具有理想的收敛精度。9个基准函数的测试结果表明:NGOA1、NGOA2的寻优性能远胜于GOA、基于曲线自适应的蝗虫优化算法(CAGOA)。NGOA2能够收敛到多个测试函数的全局最优值,是求解高维复杂函数的有效工具。

In order to solve the problem of low convergence precision and slow convergence speed of grasshopper optimization algorithm(GOA)in solving the global optimal value of high-dimensional complex function,two novel type of grasshopper optimization algorithms(NGOA)based on nonlinear weight are proposed.NGOA1 mainly improves the linear adaptive parameter in the position update formula,and uses it as the weight of the current optimal position,which is conducive to the exploration and development of the balance algorithm.On the basis of NGOA1,NGOA2 proposes an elite locust location update strategy,which not only improves the convergence speed of the algorithm,but also has ideal convergence accuracy.The test results of 9 benchmark functions show that the optimization performance of NGOA1 and NGOA2 is much better than that of GOA and algorithms based on arc adaptive grasshopper optimization algorithm(CAGOA).NGOA2 can achieve the global optimal value of multiple test functions,and is an effective tool for solving high-dimensional complex functions.