摘要
文章利用神经网络实现基本的数字逻辑功能,并在MATLAB上进行了仿真。通过加入带有噪声的输入值以测试网络的稳健性能,这是一个比较和优化不同网络性能的过程。为了找到能够更好地实现基本数字逻辑的网络模型和训练算法,文中对感知器网络、BP网络和RBF网络做出了一系列的比较分析。尤其是对BP网络中的不同训练算法,均根据MATLAB的仿真情况文中做出了比较,结果可作为网络选用依据。
In this paper,a neural network（NN） have been desgined to accomplish the function of digital logics and simulating it on the MATLAB.Then,some input with noise was given to test the its viability.It＇s a process of comparing and optimizating the performance of the different neural networks.In oder to find the better model and the training algorithm for realizing the basic digital logics.This parper makes a series comparison of the perceptron network,the BP network and the RBF network.Especially,some comparsion have been made about a few different training algorithms in BP network base on the simulation.
出处
《计算机与数字工程》
2011年第11期161-164,共4页
Computer & Digital Engineering
关键词
神经网络
数字逻辑
稳健性
网络训练
Nerural network
digital logics
viability
network training
