基于人工神经网络和遗传算法的激光器参数全局优化方法  

A Novel Global Optimization Method to Design Laser Parameters with Artificial Neural Network and Genetic Algorithm

作　　者：肖光宗[1] 龙兴武[1] 张斌[1] 吴素勇[1] 赵洪常[1] 陈文学[1] 

机构地区：[1]国防科技大学光电科学与工程学院,湖南长沙410073

出　　处：《中国激光》2010年第5期1203-1208,共6页Chinese Journal of Lasers

基　　金：国防科技大学优秀研究生创新资助项目(B090703)资助课题

摘　　要：提出了一种新颖的基于人工神经网络(ANN)和遗传算法(GA)的激光器参数全局优化方法,建立激光器输出功率的人工神经网络模型,来模拟激光器参数对输出功率的综合影响机理,进而以该模型作为目标函数,采用遗传算法对激光器参数进行全局优化。以平凹腔单横模氦氖激光器为例验证了该方法的可行性和有效性。对相同参数的激光器,人工神经网络模型的仿真数据与实验数据的均方根误差为0.0127 mW。应用该方法对其他参数全局优化后激光器预期输出功率比实验室已有的同等尺寸的激光器大,说明了该方法的有效性。A novel global optimization method based on artificial neural network and genetic algorithm is proposed for laser parameters design.An artificial neural network（ANN） is built to simulate the mechanism of the effects of laser parameters on its output power.And then the laser parameters are further optimized with genetic algorithm（GA） with the above obtained network as objective function.As an example,parameters of a typical plano-concave cavity He-Ne laser are globally optimized to testify the feasibility and effectiveness of the method.As concluded from the optimization experiment,the mean square error between the experimental data and simulation data from laser output power network is only 0.0127 mW.The output power of laser with parameters optimized by genetic algorithm is larger than the other laser with the same cavity length,which proves the effectiveness of the method.

关 键 词：激光技术 参数优化 神经网络 遗传算法 输出功率 

分 类 号：TN24[电子电信—物理电子学]