【摘要】 为了提高小波神经网络的收敛速度,文章提出了将负梯度下降法与DFP变尺度算法相结合进行权值修正的方法,在误差寻优初期采用梯度下降法迭代,当寻优过程开始接近最优时,更改寻优算法,使用DFP变尺度算法。通过仿真结果表明,改进算法减少了迭代次数,提高了算法收敛速度。更多还原

【Abstract】 To improve constringent speed of wavelet neural network,a improved weighted algorithm,which combines the method of negative gradient descent with DFP variable scale algorithm,is presented in the paper.At the initial optimization stages,iteration method based on negative gradient descent is adopted.Then DFP variable scale algorithm is adopted during the optimization process near optimum point.Finally,simulation results show that the improved weighted algorithm is effective for reducing iterations and speeding constringency.更多还原