遗传算法优化双隐含层BP神经网络（代码完整，数据齐全）

%% �ô���Ϊ�����Ŵ��㷨����������Ԥ������ % ���ջ������� clc clear close all % %% �����ṹ���� %��ȡ���� load data input output %�ڵ����� inputnum=13; hiddennum=[16 10]; outputnum=1; %ѵ�����ݺ�Ԥ������ input_train=input(1:10064,:)'; input_test=input(20065:24464,:)'; output_train=output(1:10064)'; output_test=output(20065:24464)'; % input_train=input(1:89,:)'; % input_test1=input(90:177,:)'; % n = randperm(length(input_test1)); % input_test = input_test1(:,n(1:20)); % output_train=output(1:89)'; % output_test1=output(90:177)'; % output_test = output_test1(:,n(1:20)); [inputn,inputps]=mapminmax(input_train); [outputn,outputps]=mapminmax(output_train); %�������� net=newff(inputn,outputn,[hiddennum]); %% �Ŵ��㷨������ʼ�� maxgen=20; %�������������������� sizepop=4; %��Ⱥ��ģ pcross=[0.9]; %��������ѡ����0��1֮�� pmutation=[0.05]; %��������ѡ����0��1֮�� %�ڵ����� numsum=inputnum*hiddennum(1)+hiddennum(1)+hiddennum(1)*hiddennum(2)+hiddennum(2)+hiddennum(2)*outputnum+outputnum; lenchrom=ones(1,numsum); bound=[-5*ones(numsum,1) 5*ones(numsum,1)]; %���ݷ�Χ %------------------------------------------------------��Ⱥ��ʼ��-------------------------------------------------------- individuals=struct('fitness',zeros(1,sizepop), 'chrom',[]); %����Ⱥ��Ϣ����Ϊһ���ṹ�� avgfitness=[]; %ÿһ����Ⱥ��ƽ����Ӧ�� bestfitness=[]; %ÿһ����Ⱥ��������Ӧ�� bestchrom=[]; %��Ӧ�����õ�Ⱦɫ�� %��ʼ����Ⱥ for i=1:sizepop %��������һ����Ⱥ individuals.chrom(i,:)=Code(lenchrom,bound); %���루binary��grey�ı�������Ϊһ��ʵ����float�ı�������Ϊһ��ʵ�������� x=individuals.chrom(i,:); %������Ӧ�� individuals.fitness(i)=fun(x,inputnum,hiddennum,outputnum,net,inputn,outputn); %Ⱦɫ������Ӧ�� end FitRecord=[]; %�����õ�Ⱦɫ�� [bestfitness bestindex]=min(individuals.fitness); bestchrom=individuals.chrom(bestindex,:); %���õ�Ⱦɫ�� avgfitness=sum(individuals.fitness)/sizepop; %Ⱦɫ����ƽ����Ӧ�� % ��¼ÿһ�����������õ���Ӧ�Ⱥ�ƽ����Ӧ�� trace=[avgfitness bestfitness]; %% �����������ѳ�ʼ��ֵ��Ȩֵ % ������ʼ for i=1:maxgen i % ѡ�� individuals=Select(individuals,sizepop); avgfitness=sum(individuals.fitness)/sizepop; %���� individuals.chrom=Cross(pcross,lenchrom,individuals.chrom,sizepop,bound); % ���� individuals.chrom=Mutation(pmutation,lenchrom,individuals.chrom,sizepop,i,maxgen,bound); % ������Ӧ�� for j=1:sizepop x=individuals.chrom(j,:); %���� individuals.fitness(j)=fun(x,inputnum,hiddennum,outputnum,net,inputn,outputn); end %�ҵ���С��������Ӧ�ȵ�Ⱦɫ�弰��������Ⱥ�е�λ�� [newbestfitness,newbestindex]=min(individuals.fitness); [worestfitness,worestindex]=max(individuals.fitness); % ������һ�ν��������õ�Ⱦɫ�� if bestfitness>newbestfitness bestfitness=newbestfitness; bestchrom=individuals.chrom(newbestindex,:); end individuals.chrom(worestindex,:)=bestchrom; individuals.fitness(worestindex)=bestfitness; avgfitness=sum(individuals.fitness)/sizepop; trace=[trace;avgfitness bestfitness]; %��¼ÿһ�����������õ���Ӧ�Ⱥ�ƽ����Ӧ�� FitRecord=[FitRecord;individuals.fitness]; end %% �Ŵ��㷨�������� figure(3) [r c]=size(trace); plot([1:r]',trace(:,2),'b--'); title(['Optimal individual fitness value' num2str(maxgen)]); set(gca,'fontsize',12); xlabel('Number of evolutions','fontsize',12);ylabel('fitness Value','fontsize',12); disp('��Ӧ�� ����'); %% �����ų�ʼ��ֵȨֵ��������Ԥ�� % %���Ŵ��㷨�Ż���BP��������ֵԤ�� w1=x(1:inputnum*hiddennum(1)); B1=x(inputnum*hiddennum(1)+1:inputnum*hiddennum(1)+hiddennum(1)); w2=x(inputnum*hiddennum(1)+hiddennum(1)+1:inputnum*hiddennum(1)+hiddennum(1)+hiddennum(1)*hiddennum(2)); B2=x(inputnum*hiddennum(1)+hiddennum(1)+hiddennum(1)*hiddennum(2)+1:inputnum*hiddennum(1)+hiddennum(1)+hiddennum(1)*hiddennum(2)+hiddennum(2)); w3=x(inputnum*hiddennum(1)+hiddennum(1)+hiddennum(1)*hiddennum(2)+hiddennum(2)+1:inputnum*hiddennum(1)+hiddennum(1)+hiddennum(1)*hiddennum(2)+hiddennum(2)+hiddennum(2)*outputnum); B3=x(inputnum*hiddennum(1)+hiddennum(1)+hiddennum(1)*hiddennum(2)+hiddennum(2)+hiddennum(2)*outputnum+1:inputnum*hiddennum(1)+hiddennum(1)+hiddennum(1)*hiddennum(2)+hiddennum(2)+hiddennum(2)*outputnum+outputnum); net.iw{1,1}=reshape(w1,hiddennum(1),inputnum); net.lw{2,1}=reshape(w2,hiddennum(2),hiddennum(1)); net.lw{3,2}=reshape(w3,outputnum,hiddennum(2)); net.b{1}=reshape(B1,hiddennum(1),1); net.b{2}=reshape(B2,hiddennum(2),1); net.b{3}=B3; %% BP����ѵ�� %������������ net.trainParam.epochs=1000; net.trainParam.lr=0.1; %net.trainParam.goal=0.00001; %����ѵ�� [net,per2]=train(net,inputn,outputn); %% BP����Ԥ�� %���ݹ�һ�� inputn_test=mapminmax('apply',input_test,inputps); an=sim(net,inputn_test); test_simu = mapminmax('reverse',an,outputps); error=abs(test_simu-output_test); %Ԥ��ֵ������ֵ����� errorPercent=error./output_test; %Ԥ��ֵ������ֵ�����ٷֱ� figure(1) plot(test_simu,':og','color',[0.54 0.17 0.89],'linewidth',1,'markersize',8) hold on plot(output_test,'-*','color',[0.20 0.63 0.79],'linewidth',1,'markersize',8); hold off legend('Predicted results','validation results') % % set(gca,'Fontname','arial'); % set(gca,'xtick',1:1:240);%gca是当å�?�å�? ‡è½´çš�?�å¥æŸ�?�，xtick表示我è¦è®¾ç½®x轴刻度è¦æ˜¾ç¤ºçš„ä½ç½�? % set(gca,'xticklabel',{'2-Jul','3-Jul','4-Jul','5-Jul','6-Jul','7-Jul','8-Jul','9-Jul','11-Jul','12-Jul','13-Jul','14-Jul','15-Jul','16-Jul','17-Jul','18-Jul','19-Jul','20-Jul','21-Jul','22-Jul','23-Jul','24-Jul','25-Jul','26-Jul','27-Jul','28-Jul','29-Jul','30-Jul','31-Jul','2-Aug','3-Aug','4-Aug','5-Aug','6-Aug','7-Aug','8-Aug','9-Aug','10-Aug','11-Aug','12-Aug','13-Aug','14-Aug','15-Aug','16-Aug','17-Aug','18-Aug','19-Aug','20-Aug','21-Aug','22-Aug','23-Aug','24-Aug','25-Aug','26-Aug','27-Aug','28-Aug','29-Aug','30-Aug','31-Aug','2-Sep','3-Sep','4-Sep','5-Sep','6-Sep','7-Sep','8-Sep','9-Sep','10-Sep','11-Sep','12-Sep','13-Sep','14-Sep','15-Sep','16-Sep','17-Sep','18-Sep','19-Sep','20-Sep','21-Sep','22-Sep','23-Sep','24-Sep','25-Sep','26-Sep','27-Sep','28-Sep','29-Sep','30-Sep'}); % xtickangle(90) %Ԥ������ MAPE=mean(abs((output_test - test_simu)./output_test))*100; RMSE=sqrt(mean((output_test - test_simu).^2)); figure(2) plot((output_test-test_simu)./test_simu,'-*'); title('The error of GA-BP neural network')