论文部分内容阅读
针对BP神经网络初始连接权值和阈值确定的随机性,以及网络收敛速度慢和容易陷入局部极小的问题,采用遗传算法优化BP神经网络的连接权值和阈值,构建混合GA-BPNN网络模型.利用建立的GA-BPNN模型,对湖北省宜昌地区降雨量进行插值估算,试验结果表明,单纯采用BP神经网络进行降雨量的插值估算,其归一化的平均相对误差为27.68%,而采用遗传算法优化后的BP神经网络进行降雨量插值估算,其归一化的平均相对误差为18.93%,估算的精度以及网络的稳定性和容错性都要好于单纯的BP神经网络模型.
Aiming at the randomness of determining the initial connection weight and threshold of BP neural network and the problem of slow convergence speed and easily falling into local minimum, genetic algorithm is used to optimize the connection weights and thresholds of BP neural network to construct a hybrid GA-BPNN network model By using the established GA-BPNN model, the rainfall in Yichang, Hubei Province is estimated by interpolation. The experimental results show that the average relative error of normalization is only 27.68% by using BP neural network to estimate the rainfall. The BP neural network optimized by genetic algorithm is used to estimate the rainfall interpolation. The normalized average relative error is 18.93%. The accuracy of the estimation and the stability and fault tolerance of the network are better than those of pure BP neural network model.