【摘要】 通过仿真实验发现凭经验选取小波神经网络隐层小波基,所得结构并不最优。为此,本文将遗传算法与小波神经网络结合起来进行研究。利用遗传算法来优化小波神经网络的结构,确定小波基的个数,采用BP算法来训练小波神经网络中的伸缩因子、平移因子和连接权值。仿真结果表明,该方法能准确搜索到最优结构,算法来训练小波神经网络中的伸缩因子是切实可行的。更多还原

【Abstract】 It is shown that the wavelet network structure with hidden layer decided by experience is not the optimization. In this paper we discuss the optimization to the structure of wavelet neural networks, for which the genetic algorithm is used to determine the number of elements in hidden layer while the gradient method of BP algorithm is used to compute weighting, stretching and displacing coefficients of the continuous wavelet neural networks. The effectiveness of the proposed scheme is demonstrated by examples更多还原