ex1hgf4.m

在微分控制中

close all
clear
clf reset
figure(gcf);
echo on
clc
%newff--生成一个新的bp前向神经网络
%train--训练
%sim－－仿真
pause%任意键开始
clc
%定义训练样本
%P为输入矢量
P=[-1,-2,3,1;
   -1,1,5,-3];   ;
%T为目标矢量
T=[-1,-1,1,1];
pause
clc
%创建一个新的bp前向神经网络
for j=1:1:4
    m=Z(j,1);
net=newff(minmax(P),[m,1],{'tansig','purelin'},'traingdm');
%
inputWeights=net.IW{1,1}
inputbias=net.b{1}

layerWeights=net.LW{2,1}
layerbias=net.b{2}

pause
clc
%
net.trainParam.show=50;
net.trainParam.lr=0.05;
net.trainParam.mc=0.9;
net.trainParam.epochs=1000;
net.trainParam.goal=1e-10;
pause
clc
%
tic
[net,tr]=train(net,P,T);
out=toc
t=out
pause
clc
%训练后网络权值和偏差
inputWeights=net.IW{1,1}
inputbias=net.b{1}

layerWeights=net.LW{2,1}
layerbias=net.b{2}

pause
%
A=sim(net,P)
%计算仿真误差

E=T-A
SSE=sse(E)
y=fitness(SSE,t)
end
pause
clc
echo off