为解决传统随机梯度下降(SPGD)算法参数难以实时调节、收敛速度较慢的难题,提出了一种基于自适应增益及联合指标优化的高效SPGD算法,建立了该算法的数值仿真模型,并将该算法用于千瓦级板条激光器的光束净化。仿真结果表明:与传统SPGD算法相比,提出的算法无需参数调节,且收敛速度和收敛效果均有显著提升,在对千瓦级板条激光器的光束净化实验中,激光光束质量β由7.89优化至1.95。

To solve the problems of the difficulty in adjusting parameters in real time and the convergence speed being slow in the traditional stochastic parallel gradient descent (SPGD) algorithm, this study proposes an efficient SPGD algorithm based on adaptive gain and joint index optimization and establishes a numerical simulation model of this algorithm. The proposed algorithm is used for the beam cleaning of kilowatt-class slab lasers. Simulation results show that compared with the traditional SPGD algorithm, the proposed algorithm does not require a parameter adjustment, and the convergence speed and convergence effect are significantly improved. Furthermore, in the beam purification experiment of the kilowatt-class slab laser, the laser beam quality β is optimized from 7.89 to 1.95 herein.