机构地区: 东北电力大学理学院
出 处: 《吉林大学学报(理学版)》 2010年第4期658-661,共4页
摘 要: 提出一种基于RBFNNs和PSO求解第二类Volterra积分方程的混合方法.先将积分区间离散化为点集,并代入积分方程得到方程组,再利用RBF神经网络逼近积分方程中的未知函数,将所求解问题转化为残差平方和的极小化问题.利用PSO算法求解残差平方和的极小化优化问题,得到RBF神经网络的参数,即得问题的逼近解.数值实验表明,该方法可行有效. This paper presents a hybrid method based on radial basis function neural networks(RBFNN) and particle swarm optimization(PSO) algorithm for solving the linear integral equations of the second kind of Volterra.Firstly,the integral interval is discretized into a point set.And the discretized points in the set are substituted into the equation to obtain equations.RBFNN is applied to approximating the unknown function of equations,and the solved problem can be turned into optimum problem which is solved by PSO algorithm for the advantage of PSO.Therefore,the parameters of neural networks,namely,the approximate solution,are found.Finally,numerical experiments are performed and the results show that our method is feasible.
领 域: [自动化与计算机技术] [自动化与计算机技术]