（Matlab）PSO优化（双隐层）BP神经网络算法

%% 清空环境 clc tic %读取数据 load data input output %节点个数 inputnum=2; hiddennum=5; hiddennum1=5; outputnum=1; opnum=inputnum*hiddennum+hiddennum+hiddennum*hiddennum1+hiddennum1+hiddennum1*outputnum+outputnum; % 需要优化的参数个数 %% 训练数据预测数据提取及归一化 %从1到4000间随机排序 k=rand(1,4000); [m,n]=sort(k); %划分训练数据和预测数据 input_train=input(n(1:3900),:)'; output_train=output(n(1:3900),:)'; input_test=input(n(3901:4000),:)'; output_test=output(n(3901:4000),:)'; [inputn,inputps]=mapminmax(input_train); [outputn,outputps]=mapminmax(output_train); %构建网络 net=newff(inputn,outputn,[hiddennum, hiddennum1]); % 参数初始化 %粒子群算法中的两个参数 c1 = 1.49445; c2 = 1.49445; maxgen=100; % 进化次数 sizepop=30; %种群规模 %个体和速度最大最小值 Vmax=1; Vmin=-1; popmax=5; popmin=-5; for i=1:sizepop pop(i,:)=5*rands(1,opnum); V(i,:)=rands(1,opnum); fitness(i)=H55PSOBP_fun(pop(i,:),inputnum,hiddennum,hiddennum1,outputnum,net,inputn,outputn); end % 个体极值和群体极值 [bestfitness bestindex]=min(fitness); zbest=pop(bestindex,:); %全局最佳 gbest=pop; %个体最佳 fitnessgbest=fitness; %个体最佳适应度值 fitnesszbest=bestfitness; %全局最佳适应度值 %% 迭代寻优 for i=1:maxgen i; for j=1:sizepop %速度更新 V(j,:) = V(j,:) + c1*rand*(gbest(j,:) - pop(j,:)) + c2*rand*(zbest - pop(j,:)); V(j,find(V(j,:)>Vmax))=Vmax; V(j,find(V(j,:)<Vmin))=Vmin; %种群更新 pop(j,:)=pop(j,:)+0.2*V(j,:); pop(j,find(pop(j,:)>popmax))=popmax; pop(j,find(pop(j,:)<popmin))=popmin; %自适应变异 pos=unidrnd(opnum); if rand>0.95 pop(j,pos)=5*rands(1,1); end %适应度值 fitness(j)=H55PSOBP_fun(pop(j,:),inputnum,hiddennum,hiddennum1,outputnum,net,inputn,outputn); end for j=1:sizepop %个体最优更新 if fitness(j) < fitnessgbest(j) gbest(j,:) = pop(j,:); fitnessgbest(j) = fitness(j); end %群体最优更新 if fitness(j) < fitnesszbest zbest = pop(j,:); fitnesszbest = fitness(j); end end yy(i)=fitnesszbest; end %% PSO结果分析 plot(yy) title(['适应度曲线 ' '终止代数＝' num2str(maxgen)]); xlabel('进化代数');ylabel('适应度'); x=zbest; %% 把最优初始阈值权值赋予网络预测 % %用PSO优化的BP网络进行值预测 w1=x(1:inputnum*hiddennum); B1=x(inputnum*hiddennum+1:inputnum*hiddennum+hiddennum); w2=x(inputnum*hiddennum+hiddennum+1: ... inputnum*hiddennum+hiddennum+hiddennum*hiddennum1); B2=x(inputnum*hiddennum+hiddennum+hiddennum*hiddennum1+1: ... inputnum*hiddennum+hiddennum+hiddennum*hiddennum1+hiddennum1); w3=x(inputnum*hiddennum+hiddennum+hiddennum*hiddennum1+hiddennum1+1: ... inputnum*hiddennum+hiddennum+hiddennum*hiddennum1+hiddennum1+hiddennum1*outputnum); B3=x(inputnum*hiddennum+hiddennum+hiddennum*hiddennum1+hiddennum1+hiddennum1*outputnum+1: ... inputnum*hiddennum+hiddennum+hiddennum*hiddennum1+hiddennum1+hiddennum1*outputnum+outputnum); net.iw{1,1}=reshape(w1,hiddennum,inputnum); net.lw{2,1}=reshape(w2,hiddennum1,hiddennum); net.lw{3,2}=reshape(w3,outputnum,hiddennum1); net.b{1}=reshape(B1,hiddennum,1); net.b{2}=reshape(B2,hiddennum1,1); net.b{3}=B3; %% PSO-BP网络训练 %网络进化参数 net.trainParam.epochs=120; net.trainParam.lr=0.005; net.trainParam.goal=4e-8; %网络训练 [net,per2]=train(net,inputn,outputn); %% BP网络训练 % %初始化网络结构 net1=newff(inputn,outputn,[hiddennum,hiddennum1]); % BP网络 net1.trainParam.epochs=120; net1.trainParam.lr=0.005; net1.trainParam.goal=4e-8; %网络训练 net1=train(net1,inputn,outputn); %% PSO-BP网络预测 %数据归一化 inputn_test=mapminmax('apply',input_test,inputps); inputn_train=mapminmax('apply',input_train,inputps); an=sim(net,inputn_test); an1=sim(net,inputn_train); test_PSOBP=mapminmax('reverse',an,outputps); train_PSOBP=mapminmax('reverse',an1,outputps); %% BP网络预测 %网络预测输出 an2=sim(net1,inputn_test); an3=sim(net1,inputn_train); test_BP=mapminmax('reverse',an2,outputps); train_BP=mapminmax('reverse',an3,outputps); %% PSO-BP误差输出 error_PSOBP=test_PSOBP-output_test; disp('PSO-BP results:'); errorsum_PSOBP=sum(abs(error_PSOBP)) %% PSO-BP结果绘图 figure(1); plot(test_PSOBP,':og'); hold on plot(output_test,'-*'); legend('Predictive output','Expected output','fontsize',10.8); title('PSO-BP network output','fontsize',12); xlabel("samples",'fontsize',12); figure(2); plot(error_PSOBP,'-*'); title('PSO-BP Neural network prediction error'); xlabel("samples",'fontsize',12); figure(3); plot(100*(output_test-test_PSOBP)./output_test,'-*'); title('PSO-BP Neural network prediction error percentage (%)'); xlabel("samples",'fontsize',12); figure(4); plot(100*(output_train-train_PSOBP)./output_train,'-*'); title('PSO-BP Neural network training error percentage (%)'); xlabel("samples",'fontsize',12); %% BP误差输出 error_BP=test_BP-output_test; disp('BP results:'); errorsum_BP=sum(abs(error_BP)) %% BP结果绘图 figure(5); plot(test_BP,':og'); hold on plot(output_test,'-*'); legend('Predictive output','Expected output','fontsize',10.8); title('BP network output','fontsize',12); xlabel("samples",'fontsize',12); figure(6); plot(error_BP,'-*'); title('BP Neural network prediction error'); xlabel("samples",'fontsize',12); figure(7); plot(100*(output_test-test_BP)./output_test,'-*'); title('BP Neural network prediction error percentage (%)'); xlabel("samples",'fontsize',12); figure(8); plot(100*(output_train-train_BP)./output_train,'-*'); title('BP Neural network training error percentage (%)'); xlabel("samples",'fontsize',12); toc