Research on improved Rider Optimization Algorithm based on kernel density estimation
Huanzhang Ling, Xufeng Tan, Chenglei Gu, Hongjiao Bi, Shuang Liu · Research Square · 2023
Abstract Combining the randomness of the solution convergence process due to the random search of the solution space by the Estimation of the Distribution Algorithm and the characteristics of fast convergence and high accuracy of the Rider Optimization Algorithm, a hybrid Estimation of Distribution Algorithm combining them is proposed. In order to further accelerate the convergence speed of the algorithm, the strategy adopted by the original Rider Optimization Algorithm is adjusted. The kernel density estimation is used to properly select the iteration cycle of the two algorithms, the bandwidth adopted by the Estimation of Distribution Algorithm and other relevant parameters, and the Estimation of Distribution Algorithm is used to adjust the historical optimal position and maximum speed required in the Rider Optimization Algorithm. The numerical results of benchmark test functions show that the proposed optimization algorithm can achieve a certain balance between exploring the global optimal solution and characterizing the local optimal solution, and can quickly converge to the global optimal solution of complex optimization problems.