基于PCA-GABP神经网络的发动机负荷特性预测

针对BP算法存在的不足,本文提出了一种PCA-GABP神经网络方法预测发动机负荷特性,该方法由主成分分析(PCA)和遗传神经网络(GABP)两部分构成,采用PCA技术减少网络输入变量、精简网络结构、提高学习效率;GABP算法采用局部改进遗传算法优化神经网络权值,并采用自适应学习速率动量梯度下降算法对神经网络进行训练。预测结果表明该方法在准确性和收敛性方面都优于BP算法。

Study on Prediction of Engine Load Characteristic Based on PCA-GABP Neural Network

With regard to the drawbacks of BP algorithm,in this paper a PCA-GABP neural network method for engine load characteristic prediction is presented.The method consists of two parts: Principal Component Analysis(PCA) and Genetic Algorithms Back Propagation(GABP) algorithm.The PCA technology is used to reduce input variables,simplifying the network structure and improving the learning efficiency;The GABP algorithm is introduced through partially improved genetic algorithm to optimize the weights of the neural network,and use adaptive learning rate momentum gradient descent algorithm to train neural network.The results show that the method has a better accuracy and convergence than BP algorithm.