【神经网络解耦】基于matlab的神经网络解耦系统仿真

1.软件版本

matlab2013b

2.本算法理论知识

3.核心代码

clc
clear
close all;
warning off;
addpath 'functions'
 

%外部输入的标准的值，这里可以在具体的验证的时候加入扰动。
CR   = 0.4;
h    = 0.6;
selt = 1;%1:加入扰动；0：不加扰动



%以下两个值越大，那么其PSO优化的RBF性能就越好
%进化次数  
iteration  = 250;
%种群规模
Sizes      = 20;   

sel        = 0;%1进行PSO优化得到最佳值,0直接进行实际测试

%本代码是在普通的PSO下的RBF神经网络解耦程序
%本代码是在普通的PSO下的RBF神经网络解耦程序
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%以下为PSO%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
if sel == 1
   %输入的CR和h
   %分别仅输入CR和h，使其达到解耦的结果
   [yy,Zbest] = func_train_online(iteration,Sizes,CR,h);
   figure(1)
   plot(yy,'LineWidth',2);grid on;
   xlabel('进化代数');
   ylabel('适应度');
   individual=Zbest;
   save trainPSO.mat Zbest yy
else
   load trainPSO.mat
   figure(1)
   plot(yy,'LineWidth',2);grid on;
   xlabel('进化代数');
   ylabel('适应度');
   individual=Zbest; 
end


 
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%以下为RBF%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
w11=reshape(individual(1:6),3,2);
w12=reshape(individual(7:12),3,2);
w13=reshape(individual(13:18),3,2);

w21=individual(19:27);
w22=individual(28:36);
w23=individual(37:45);

rate1=0.006;rate2=0.001; %学习率
k=0.3;K=3;
y_1=zeros(3,1);y_2=y_1;y_3=y_2;   %输出值
u_1=zeros(3,1);u_2=u_1;u_3=u_2;   %控制率
h1i=zeros(3,1);h1i_1=h1i;  %第一个控制量
h2i=zeros(3,1);h2i_1=h2i;  %第二个控制量
h3i=zeros(3,1);h3i_1=h3i;  %第三个空置量
x1i=zeros(3,1);x2i=x1i;x3i=x2i;x1i_1=x1i;x2i_1=x2i;x3i_1=x3i;   %隐含层输出 

%权值初始化
k0=0.03;

%值限定
ynmax=1;ynmin=-1;  %系统输出值限定
xpmax=1;xpmin=-1;  %P节点输出限定
qimax=1;qimin=-1;  %I节点输出限定
qdmax=1;qdmin=-1;  %D节点输出限定
uhmax=1;uhmin=-1;  %输出结果限定

for k=1:1:6000
    k
    %系统输出
    y1(k) = (0.4*y_1(1)+u_1(1)/(1+u_1(1)^2)+0.2*u_1(1)^3+0.5*u_1(2))+0.3*y_1(2);
    y2(k) = (0.4*y_1(2)+u_1(2)/(1+u_1(2)^2)+0.2*u_1(2)^3+0.5*u_1(1))+0.3*y_1(1);
    y3(k) = 0;
    
    
    %控制目标
    if selt == 1%加扰测试
        r1(k)  = CR + 0.005*sin(2*pi*k/200);
        r2(k)  = h  + 0.015*sin(2*pi*k/200);
        r3(k)  = 0; 
        r1s(k) = CR;
        r2s(k) = h;
        r3s(k) = 0;         
    else        %跟踪测试
        r1(k)  = sign(0.001*sin(2*pi*k/2000));
        r2(k)  = sign(0.003*sin(2*pi*k/2000));
        r3(k)  = 0;      
        r1s(k) = CR;
        r2s(k) = h;
        r3s(k) = 0;    
    end
    
    
    %系统输出限制
    yn=[y1(k),y2(k),y3(k)];
    yn(find(yn>ynmax))=ynmax;
    yn(find(yn<ynmin))=ynmin;
    
    %输入层输出
    x1o=[r1(k);yn(1)];
    
    x2o=[r2(k);yn(2)];
    
    x3o=[r3(k);yn(3)];
    
    %隐含层 
    x1i=w11*x1o;
    x2i=w12*x2o;
    x3i=w13*x3o;

    %比例神经元P计算
    xp=[x1i(1),x2i(1),x3i(1)];
    xp(find(xp>xpmax))=xpmax;
    xp(find(xp<xpmin))=xpmin;
    qp=xp;
    h1i(1)=qp(1);
    h2i(1)=qp(2);
    h3i(1)=qp(3);

    %积分神经元I计算
    xi=[x1i(2),x2i(2),x3i(2)];
    qi=[0,0,0];
    qi_1=[h1i(2),h2i(2),h3i(2)];
    
    qi=qi_1+xi;
    qi(find(qi>qimax))=qimax;
    qi(find(qi<qimin))=qimin;
    h1i(2)=qi(1);
    h2i(2)=qi(2);
    h3i(2)=qi(3);

    %微分神经元D计算
    xd=[x1i(3),x2i(3),x3i(3)];
    qd=[0 0 0];
    xd_1=[x1i_1(3),x2i_1(3),x3i_1(3)];
    qd=xd-xd_1;
    qd(find(qd>qdmax))=qdmax;
    qd(find(qd<qdmin))=qdmin;
    h1i(3)=qd(1);
    h2i(3)=qd(2);
    h3i(3)=qd(3);

    %输出层计算
    wo=[w21;w22;w23];
    qo=[h1i',h2i',h3i'];
    qo=qo';
    uh=wo*qo;
    uh(find(uh>uhmax))=uhmax;
    uh(find(uh<uhmin))=uhmin;
    u1(k)=uh(1);
    u2(k)=uh(2);
    u3(k)=uh(3);  

    
    %计算误差
    error=[r1(k)-y1(k);r2(k)-y2(k);0];  
    error1(k)=error(1);error2(k)=error(2);error3(k)=0;
    J(k)=0.5*(error(1)^2+error(2)^2);   %调整大小
    ypc=[y1(k)-y_1(1);y2(k)-y_1(2);y3(k)-y_1(3)];
    uhc=[u_1(1)-u_2(1);u_1(2)-u_2(2);u_1(3)-u_2(3)];
    
    %隐含层和输出层权值调整

    %调整w21
    Sig1=sign(ypc./(uhc(1)+0.00001));
    dw21=sum(error.*Sig1)*qo';  
    w21=w21+rate2*dw21;
    
    %调整w22
    Sig2=sign(ypc./(uh(2)+0.00001));
    dw22=sum(error.*Sig2)*qo';
    w22=w22+rate2*dw22;
    
    %调整w23
    Sig3=sign(ypc./(uh(3)+0.00001));
    dw23=sum(error.*Sig3)*qo';
    w23=w23+rate2*dw23;

    %输入层和隐含层权值调整
    delta2=zeros(3,3);
    wshi=[w21;w22;w23];
    for t=1:1:3
        delta2(1:3,t)=error(1:3).*sign(ypc(1:3)./(uhc(t)+0.00000001));
    end
    for j=1:1:3
        sgn(j)=sign((h1i(j)-h1i_1(j))/(x1i(j)-x1i_1(j)+0.00001));
    end
 
     s1=sgn'*[r1(k),y1(k)];
     wshi2_1=wshi(1:3,1:3);
     alter=zeros(3,1);
     dws1=zeros(3,2);
     for j=1:1:3
         for p=1:1:3
             alter(j)=alter(j)+delta2(p,:)*wshi2_1(:,j);
         end
     end
     
     for p=1:1:3
         dws1(p,:)=alter(p)*s1(p,:);
     end
     w11=w11+rate1*dws1;

     %调整w12
    for j=1:1:3
        sgn(j)=sign((h2i(j)-h2i_1(j))/(x2i(j)-x2i_1(j)+0.0000001));
    end
    s2=sgn'*[r2(k),y2(k)];
    wshi2_2=wshi(:,4:6);
    alter2=zeros(3,1);
    dws2=zeros(3,2);
    for j=1:1:3
        for p=1:1:3
            alter2(j)=alter2(j)+delta2(p,:)*wshi2_2(:,j);
        end
    end
    for p=1:1:3
        dws2(p,:)=alter2(p)*s2(p,:);
    end
    w12=w12+rate1*dws2;
    
    %调整w13
    for j=1:1:3
        sgn(j)=sign((h3i(j)-h3i_1(j))/(x3i(j)-x3i_1(j)+0.0000001));
    end
    s3=sgn'*[r3(k),y3(k)];
    wshi2_3=wshi(:,7:9);
    alter3=zeros(3,1);
    dws3=zeros(3,2);
    for j=1:1:3
        for p=1:1:3
            alter3(j)=(alter3(j)+delta2(p,:)*wshi2_3(:,j));
        end
    end
    for p=1:1:3
        dws3(p,:)=alter2(p)*s3(p,:);
    end
    w13=w13+rate1*dws3;

    %参数更新
    u_3=u_2;u_2=u_1;u_1=uh;
    y_2=y_1;y_1=yn;
    h1i_1=h1i;h2i_1=h2i;h3i_1=h3i;
    x1i_1=x1i;x2i_1=x2i;x3i_1=x3i;
    
    ErrCr(k) = y1(k) - r1s(k);
    Errh(k)  = y2(k) - r2s(k);
end

 

if selt == 0
    time=0.001*(1:k);
    figure(2)
    subplot(211)
    plot(time,r1,'r-',time,y1,'b-');
    ylabel('CR');
    legend('控制目标','实际输出','fontsize',12);
    subplot(212)
    plot(time,r2,'r-',time,y2,'b-');
    ylabel('h');
    legend('控制目标','实际输出','fontsize',12);
else
    time=0.001*(1:round(k)/20);
    figure(3)
    subplot(211)
    plot(time,ErrCr(1:round(k)/20),'k-','LineWidth',2);
    title('CR输入输出误差');
    axis([0,time(round(k)/20),-0.8,0.4]);

    
    subplot(212)
    plot(time,Errh(1:round(k)/20),'k-','LineWidth',2);
    title('h输入输出误差');
    axis([0,time(round(k)/20),-0.8,0.4]);
    
    save errorpso.mat ErrCr Errh
    
    
end

4.操作步骤与仿真结论

普通PSO下的RBF解耦仿真结果：

这个PSO粒子群的优化适应度曲线，其反应了PSO的收敛情况以及最终的性能。

这个是系统的跟踪效果。

加入扰动之后的误差曲线。

HPSO下的RBF解耦仿真结果：

HPSO的收敛曲线。

HPSO的跟踪效果。

HPSO的加入扰动后的误差曲线。

5.参考文献

[1]付龙海, 李蒙. 基于PID神经网络解耦控制的变风量空调系统[J]. 西南交通大学学报, 2005, 40(1):13-17.

A05-05