国产欧美精品一区二区,中文字幕专区在线亚洲,国产精品美女网站在线观看,艾秋果冻传媒2021精品,在线免费一区二区,久久久久久青草大香综合精品,日韩美aaa特级毛片,欧美成人精品午夜免费影视

PID neural network (PIDNN) is a neural network model that integrates proportional, differential and integral links, and has simple structure . It is suitable for nonlinear system identification. However, the BP algorithm used in the network is sensitive to the initial weight and sample quality, and it is difficult to select the parameters, which leads to the slow convergence of the network and easy to fall into the local minimum. A method of optimizing network weight by cultural genetic algorithm is proposed. Differential evolution algorithm in global optimization based on the results, the chaotic local search algorithm, and further optimize the network weights; according to the prior knowledge, using L1 regularization, the objective function of the regularization algorithm to search, to avoid potential solutions, ensure the generalization ability of network model. A hybrid nonlinear system is identified and simulated. The simulation results show that the optimized neural network has high recognition precision and good generalization ability.