两类新型神经网络及其在安全评价中的应用  

Two Types of New Neural Networks and Their Applications to Safety Assessment

作　　者：郭宇航[1] 王保国[1] 

机构地区：[1]北京理工大学宇航科学技术学院,北京100081

出　　处：《中国安全科学学报》2008年第7期29-33,共5页China Safety Science Journal

基　　金：国家自然科学基金资助(50376004);高等学校博士学科点专项基金资助(20030007028)

摘　　要：详细讨论小波神经网络(WNN)和模糊神经网络(FNN)的构造以及训练学习过程,并针对安全评价问题分别完成算例。典型算例表明:小波神经网络具有很好的逼近与映射能力,并且有很强的泛化能力;模糊神经网络将模糊数学与人工神经网络相互融合起来,有效提升了系统的智能功能。两类新型神经网络使得人-机-环境系统工程中的许多安全评价问题有了更广泛的量化工具,并具有安全评价的量化较准确的特点。The structure and training learning process about wavelet neural networks and fuzzy neural network are discussed in detail. Applied examples of the two types of new neural networks in safety assessment are given. It is concluded that wavelet neural network has good capabilities of approximating, mapping and generalizing; and fuzzy neural network effectively upgrades the system＇s intelligence function due to combining fuzzy mathematics with artificial neural networks. The two types of new neural networks offer more extensive quantification tools for many safety assessment problems in human-machine-environment system engineering, and make the quantification of safety assessment more accurate.

关 键 词：小波神经网络(WNN) 模糊神经网络(FNN) 安全评价 人机与环境 安全工程 

分 类 号：X913.4[环境科学与工程—安全科学]