7-2-1.m

《神经网络理论与Matlab R2007实现》（程序代码）

x1=rands(400)*pi;
x2=rands(300)*pi;
x3=zeros(100,1)*pi;
p1=x1;
p2=[x2;x3];
P=[p1 p2]';
deltatt=0.05;
T'=p2+deltatt*(10*sin(p1)-2*p2);
net=newff(minmax(P),[9 1],{'tansig','purelin'});
net.trainParam.epochs=2000;
net=train(net,P,T);
%利用x2=0的测试样本求网络的仿真输出
a=0:0.1:3.9;
b=zeros(1,40);
P_test=[a;b];
y=sim(net,P_test);
nf=0.5*sin(a);
plot(a,y);
hold on
plot(a,nf,'*');
hold off
figure;
%利用x1=π/2的测试样本求网络的仿真输出
a=ones(1,40)*0.5*pi;
b=-10:0.5:9.5;
P_test=[a;b];
y=sim(net,P_test);
nf=0.5+0.9*b;
plot(b,y);
hold on
plot(b,nf,'*');
hold off