基于神经网络代数算法的电磁环境评价  

The Electromagnetic Environment Evaluation Based on Algebra Algorithm of Neural Network

作　　者：黄蓉[1] 宋振明[1] 

机构地区：[1]西南交通大学数学学院,四川成都610031

出　　处：《洛阳师范学院学报》2013年第11期47-50,共4页Journal of Luoyang Normal University

基　　金：国家自然科学基金(61175055);四川省科技支撑计划(2011FZ0051)

摘　　要：前馈三层神经网络具有很好的非线性映射能力,而前馈三层神经网络的代数算法能够实现零代价函数的网络精确学习过程,本文将其应用于电磁环境复杂度的评价.首先从时域、空域、频域和能量域四个方面选择电磁环境的评价指标,再建立零代价函数的电磁环境评价模型,进一步说明前馈三层神经网络的代数算法在电磁环境评价上的应用.Feed forward three-layer neural network has good ability of nonlinear mapping, and algebra algo- rithm of neural network can achieve accurate network learning process with zero cost function, which can be applied to evaluate electromagnetic environment complexity. First, evaluation indexes could be chosen from time domain, airspace, frequency domain and the energy domain. Then electromagnetic environment evaluation model with zero cost function is established. This model can further explain the application of algebra algorithm of neural network on electromagnetic environmental evaluation.

关 键 词：电磁环境 零代价函数 代数算法 评价指标 

分 类 号：O441[理学—电磁学]