|
| 摘要: |
| 本文研究了大规模线性方程问题的扰动Hopfield神经网络, 并给出了网络收敛的判别准则. 在一定条件下, 网络的稳态误差一致有界或者收敛于0, 网络具有较好的鲁棒性. 最后数值仿真验证了方法的有效性. |
| 关键词: 大规模联立线性方程 扰动梯度神经网络 系统的鲁棒性 |
| DOI： |
| 分类号:O233 |
| 基金项目:Supported by the Teaching Reform of Higher Vocational Education and Re-search Project of Zhengzhou University of Light Industry |
|
| THE REAL-TIME SOLUTION OF LARGE-SCALE SIMULTANEOUS LINEAR EQUATIONS BASED ON NEURAL NETWORK |
|
ZHAO Yan,ZHAO Wei-feng,LIAO Wu-dai
|
| Abstract: |
| The paper studies the disturbed Hopfield neural network of the large-scale simul-taneous linear equations, and points out the criterion of neural network convergence. Under certain conditions, the steady-state errors of this network are consistently bounded or convergent to zero. That is to say, the network system has good robustness. We give a simulation example, and verify the authenticity of the conclusion. |
| Key words: the large-scale simultaneous linear equation disturbed gradient neural network robustness |
