安徽工业大学管理科学与工程学院基于MATLAB神经网络仿真摘要随着人工神经网络的研究和应用越来越广泛，误差反向传播算法（BP算法）的提出，成功地解决了求解非线性连续函数的多层前馈神经网络权值调整问题，BP神经网络如今成为最广泛使用的网络，研究它对探索非线性复杂问题具有重要意义，而且它具有广泛的应用前景。以BP神经网络为例，讨论了BP神经网络及几种改进BP神经网络性能的算法；通过BP学习算法的推导和分析得知BP网络是一种多层前馈网络，采用最小均方差的学习方式，缺点是仅为有导师训练，训练时间长，易限于局部极小；运用MATLAB来实现各种BP神经网络的实现的设计与训练，比较不同BP神经网络的性能，验证改进BP网络的优势，得出如何根据对象选取神经网络的结论。关键词：人工神经网络、BP神经网络、误差反向传播算法、MATLAB、仿真共45页第1页┊┊┊┊┊┊┊┊┊┊┊┊┊装┊┊┊┊┊订┊┊┊┊┊线┊┊┊┊┊┊┊┊┊┊┊┊┊安徽工业大学管理科学与工程学院基于MATLAB神经网络仿真AbstractWiththeartificialneuralnetworkofresearchandapplicationofmoreandmorewidely,theerrorback-propagationalgorithm(BPalgorithm)isproposed,successfullyresolvedthecontinuousfunctionforsolvingnonlinearmulti-layerfeed-forwardneuralnetworkweightsadjustment,BPnetworkhasbecomenowthemostwidelyusednetworks,Studytoexploreitscomplicatednonlinearproblemhasimportantsignificance,butalsohasbroadapplicationprospects.BPneuralnetworkisdiscussedandseveralimprovementsintheperformanceofBPneuralnetworkalgorithm.BPlearningalgorithmthroughthederivationandanalysisthattheBPnetworkisamulti-layerfeedforwardnetworks,theuseofleast-mean-varianceapproachtolearning,thereisonlydisadvantageisthatthetraininginstructors,trainingtime,limitedtolocalminimumeasily.TheuseofMATLABtoachieveavarietyofBPneuralnetworktoachievethedesignandtraining,tocomparetheperformanceofBPneuralnetworktoverifytheadvantagesofimprovingtheBPnetwork,howtodrawtheobjectselectedinaccordancewiththeconclusionsofneuralnetworks.Keywords:Artificialneuralnetwork,BPneuralnetworks,errorback-propagationalgorithm,MATLAB,simulation共45页第2页┊┊┊┊┊┊┊┊┊┊┊┊┊装┊┊┊┊┊订┊┊┊┊┊线┊┊┊┊┊┊┊┊┊┊┊┊┊安徽工业大学管理科学与工程学院基于MATLAB神经网络仿真目录1.绪论..................................................................51.1引言..............................................................51.2神经网络概述......................................................51．2．1神经网络起源................................................51．2．2神经网络的发展历程..........................................51．2．3神经网络国内发展概况........................................61．2．4神经网络研究现状............................................71.3研究目的、方法和问题（BP神经网络）................................81．3．1研究目的....................................................81．3．2研究方法....................................................81．3．3研究问题....................................................82.BP神经网络..........................................................102.1BP神经网络相关原理...............................................102．1．1神经元非线性模型...........................................102．1．2有教师监督学习.............................................102．1．3神经元数学模型.............................................112．1．4Delta学习规则..............................................112．1．5神经元激活函数.............................................122．1．6BP神经网络收敛准则......................
