吉林大学学报(理学版) ›› 2020, Vol. 58 ›› Issue (2): 277-284.

• 数学 • 上一篇    下一篇

基于Chebyshev神经网络的非线性Fredholm积分方程数值解法

李娜, 韩惠丽, 房彦兵   

  1. 宁夏大学 数学统计学院, 银川 750021
  • 收稿日期:2019-06-21 出版日期:2020-03-26 发布日期:2020-03-25
  • 通讯作者: 韩惠丽 E-mail:nxhan@ 126.com

Numerical Algorithm of Nonlinear Fredholm IntegralEquation Based on Chebyshev Neural Network

LI Na, HAN Huili, FANG Yanbing   

  1. School of Mathematics and Statistics, Ningxia University, Yinchuan 750021, China
  • Received:2019-06-21 Online:2020-03-26 Published:2020-03-25
  • Contact: HAN Huili E-mail:nxhan@ 126.com

摘要: 利用Chebyshev扩展块代替隐层结构, 提出一种基于函数逼近的Chebyshev神经网络模型求解非线性Fredholm积分方程的方法, 并给出其最佳逼近解及算法的收敛性分析. 数值算例验证了算法的可行性和有效性.

关键词: Chebyshev多项式, 神经网络, Fredholm积分方程

Abstract: Using the Chebyshev expansion block instead of  the hidden layer structure, we proposed a method of Chebyshev neural network model based on function approximation to solve nonlinear Fredholm integral equations, and gave its best  approximation solution and the convergence analysis of the algorithm. The feasibility and effectiveness of the algorithm were verified by numerical examples.

Key words: Chebyshev polynomial, neural network, Fredholm integral equation

中图分类号: 

  • O175.5