光线寻优算法在函数优化中的应用

Light Ray Optimization Algorithm and its Applications to Function Optimization

光线寻优算法是一种模拟光传播过程的智能优化算法,具有可调参数少、结构简单、容易实现等优点。该算法用网格划分可行域,将具有不同折射率的介质填充到各网格中,并将光在此变折射率介质中的传播路径设想成算法的寻优路径,从而达到自动搜索寻优的目的。将光线寻优算法用于求解文献中的6个标准测试函数,并与模拟退火算法、保留精英遗传算法、标准粒子群算法进行比较,通过大量数值实验验证了算法的可行性、有效性及潜在的优越性。

The light ray optimization algorithm as an intelligent optimization algorithm simulating the propagation process of light rays is discussed, which has the advantages of simple structure, few tuning parameters and easy tuning. The feasible region is divided by grids in which the media with different refracitivities are put, and then a beam of light rays propagate in these media to search the optimal val- ue automatically in this algorithm. The light ray optimization is used to solve six standard test functions in literature, and is compared with the simulated annealing, the elitist genetic algorithm, and the standard particle swarm optimization. The numerical experiments show the feasibility, validity and potential advantages of the proposed algorithm.