PSO优化BP网络:主函数:%清空环境变量clc
clear
%
%%网络结构建立%读取数据load data input output
%节点个数inputnum=2;
hiddennum=5;
outputnum=1;
%训练数据和预测数据input_train=input(1:1900,:)';
input_test=input(1901:2000,:)';
output_train=output(1:1900)';
output_test=output(1901:2000)';
%选连样本输入输出数据归一化[inputn,inputps]=mapminmax(input_train);
[outputn,outputps]=mapminmax(output_train);
%构建网络net=newff(inputn,outputn,hiddennum);
net.trainParam.epochs=500;
net.trainParam.lr=0.1;
net.trainParam.goal=0.00001;
%net.trainParam.show=NaN
%网络训练net=train(net,inputn,outputn);
%数据归一化inputnbp_test = mapminmax('apply',input_test,inputps);
an=sim(net,inputnbp_test);
testout_simu=mapminmax('reverse',an,outputps);
%网络预测结果图形figure(1)
plot(testout_simu,'og');
hold on
plot(output_test,'-*');
legend('预测输出','期望输出');
title('BP网络预测输出','fontsize',12);
xlabel('样本','fontsize',12);ylabel('函数输出','fontsize',12);
%网络预测模型error1=testout_simu-output_test;
%%PSO算法参数初始化maxgen=50; %进化代数,即迭代次数sizepop=30; %种群规模c1=1.49445; %学习因子c2=1.49445; %学习因子
%节点总数numsum=inputnum*hiddennum+hiddennum+hiddennum*outputnum+outputnum;
%个体和速度最大最小值popmax=5;
popmin=-5;
vmax=1;
vmin=-1;
%------------------------------------------------------种群初始化--------------------------------------------------------
fori=1:sizepop
%随机产生一个种群pop(i,:)=5*randn(1,21); %初始化粒子v(i,:)=randn(1,21); %初始化速度x=pop(i,:)
%计算粒子适应度值fitness(i)=fun(x,inputnum,hiddennum,outputnum,net,inputn,outputn);
end
%寻找初始极值[bestfitness bestindex]=min(fitness);
zbest=pop(bestindex,:); %群体极值位置gbest=pop; %个体极值位置fitnessgbest=fitness;%个体极值适应度值fitnesszbest=bestfitness;%群体极值适应度值%%迭代求解最佳初始阀值和权值%迭代寻优for i=1:maxgen
%粒子位置和速度更新for j=1:sizepop
%速度更新v(j,:)=v(j,:)+c1*rands*(gbest(j,:)-pop(j,:))+c2*rand*(zbest-pop(j,:));
v(j,find(v(j,:)>vmax))=vmax;
v(j,find(v(j,:)
%位置更新pop(j,:)=pop(j,:)+0.5*v(j,:);
pop(j,find(pop(j,:)>popmax))=popmax;
pop(j,find(pop(j,:)
%新粒子适应度值fitness(j)=fun(x,inputnum,hiddennum,outputnum,net,inputn,outputn);
end
%个体极值和群体极值更新for j=1:sizepop
%个体极值更新if fitness(j)
zbest=pop(j,:);
fitnesszbest=fitness(j);
end
%群体极值更新if fitness(j)
zbest=pop(j,:);
fitnesszbest=fitness(j);
end
end
%每代最优值记录到yy数组中yy(i)=fitnesszbest;
end
%%PSO算法结果分析figure(2)
plot(yy);
title('最优个体适应度值','fontsize',12);
xlabel('进化代数','fontsize',12);ylabel('适应度','fontsize',12);
%%把最优初始阀值权值赋予网络预测% %用遗传算法优化的BP网络进行值预测x=fitnesszbest;
w1=x(1:inputnum*hiddennum);
B1=x(inputnum*hiddennum+1:inputnum*hiddennum+hiddennum);
w2=x(inputnum*hiddennum+hiddennum+1:inputnum*hiddennum+hiddennum+hiddennum*outputnum);
B2=x(inputnum*hiddennum+hiddennum+hiddennum*outputnum+1:inputnum*hiddennum+hiddennum+hiddennum*outputnum+outputnum);
net.iw{1,1}=reshape(w1,hiddennum,inputnum);
net.lw{2,1}=reshape(w2,outputnum,hiddennum);
net.b{1}=reshape(B1,hiddennum,1);
net.b{2}=B2;
%%BP网络训练%网络进化参数net.trainParam.epochs=500;
net.trainParam.lr=0.1;
%net.trainParam.goal=0.00001;
%网络训练[net,per2]=train(net,inputn,outputn);
%%BP网络预测%数据归一化inputn_test=mapminmax('apply',input_test,inputps);
an=sim(net,inputn_test);
test_simu=mapminmax('reverse',an,outputps);
figure(3)
plot(test_simu,'og');
hold on
plot(output_test,'-*');
legend('预测输出','期望输出');
title('PSOBP网络预测输出','fontsize',12);
xlabel('样本','fontsize',12);ylabel('函数输出','fontsize',12);
error=test_simu-output_test;
figure(4)
plot(error1,'--og');
hold on
plot(error,'-*');
legend('BP预测误差','PSOBP预测误差');
title('预测误差','fontsize',12);
xlabel('样本','fontsize',12);ylabel('误差','fontsize',12);
适应度函数:function error = fun(x,inputnum,hiddennum,outputnum,net,inputn,outputn)
%该函数用来计算适应度值%x input输入粒子位置%inputnum input输入层节点数%outputnum input隐含层节点数%net input网络%inputn input训练输入数据%outputn input训练输出数据
%error output个体适应度值
%提取w1=x(1:inputnum*hiddennum);
B1=x(inputnum*hiddennum+1:inputnum*hiddennum+hiddennum);
w2=x(inputnum*hiddennum+hiddennum+1:inputnum*hiddennum+hiddennum+hiddennum*outputnum);
B2=x(inputnum*hiddennum+hiddennum+hiddennum*outputnum+1:inputnum*hiddennum+hiddennum+hiddennum*outputnum+outputnum);
net=newff(inputn,outputn,hiddennum);
%网络进化参数net.trainParam.epochs=20;
net.trainParam.lr=0.1;
net.trainParam.goal=0.00001;
net.trainParam.show=100;
net.trainParam.showWindow=0;
%网络权值赋值net.iw{1,1}=reshape(w1,hiddennum,inputnum);
net.lw{2,1}=reshape(w2,outputnum,hiddennum);
net.b{1}=reshape(B1,hiddennum,1);
net.b{2}=B2;
%网络训练net=train(net,inputn,outputn);
an=sim(net,inputn);
error=sum(abs(an-outputn));运行程序出错:
Error using rands (line 28)
Not enough input arguments.
Error in psobp (line 94)
v(j,:)=v(j,:)+c1*rands*(gbest(j,:)-pop(j,:))+c2*rands*(zbest-pop(j,:));
请高手帮忙解决一下。谢谢!