【BP数据预测】基于matlab差分进化算法优化BP神经网络DE-BP数据预测【含Matlab源码 1315期】

%% 差分进化算法应用于优化BP神经网络的初始权值和阈值 %% 清空环境变量 clear all; clc; warning off load v357; load y357; Pn_train=v; Tn_train=y; Pn_test=v; Tn_test=y; P_train=v; T_train=y; % P_train=[0 25.27 44 62.72 81.4 100.2; % 290.5 268.8 247.2 224.5 206 184.4; % 0 16.12 33.25 50.42 67.62 84.73; % 542.5 517.8 493 465.3 435.6 410.8; % 0 11.1 28.1 44.93 61.38 78.57; % 826.1 800.2 769.1 740.0 706.2 669.3]; % T_train=[0 1 2 3 4 5];%以上是未处理的数据 % P_test=[0 25.25 43 62.75 81.6 100.7; % 290.3 268.4 247.5 224.6 206 184.2; % 0 16.14 33.26 50.47 67.68 84.79; % 542.7 517.9 495 465.8 435.6 410.9; % 0 11.4 28.6 44.94 61.36 78.59; % 826.3 800.7 769.8 740.5 706.7 669.3]; % T_test=[0 1 2 3 4 5]; % Pn_train=[0 0.252 0.439 0.626 0.813 1 0 0.19 0.392 0.595 0.798 1 0 0.141 0.358 0.572 0.781 1; % 1 0.795 0.592 0.378 0.204 0 1 0.815 0.626 0.415 0.189 0 1 0.835 0.637 0.451 0.235 0]; % %T 为目标矢量 ，归一化后的数据 % Tn_train=[0.05,0.23,0.41,0.59,0.77,0.95,0.05,0.23,0.41,0.59,0.77,0.95,0.05,0.23,0.41,0.59,0.77,0.95]; % Pn_test=[ 0 0.17 0.39 0.595 0.798 1 0 0.141 0.358 0.572 0.781 1 0 0.258 0.439 0.626 0.813 1; % 1 0.815 0.625 0.415 0.189 0 1 0.835 0.635 0.451 0.235 0 1 0.795 0.599 0.378 0.204 0 ]; % Tn_test=[0.05,0.23,0.41,0.59,0.77,0.95,0.05,0.23,0.41,0.59,0.77,0.95,0.05,0.23,0.41,0.59,0.77,0.95]; %% 参数设置 S1 = size(Pn_train,1); % 输入层神经元个数 S2 = 6; % 隐含层神经元个数 S3 = size(Tn_train,1); % 输出层神经元个数 Gm=10; %最大迭代次数 F0=0.5; %F为缩放因子 Np=5; %种群规模 CR=0.5; %杂交参数 G=1;%初始化代数 N=S1*S2 + S2*S3 + S2 + S3;%所求问题的维数 %% 创建BP神经网络 %net_optimized = newff(Pn_train,Tn_train,S2); net_optimized=newff(minmax(Pn_train),[6,1],{'logsig','purelin'},'traingdm');%隐含层神经元S型正切,输出层S型对数，动量梯度下降法训练BP网络, %%差分进化算法 ge=zeros(1,Np);%各代的最优值 bestx=zeros(Np,N);%各代的最优解 %产生初始种群 xmin=-1;xmax=1; X0=(xmax-xmin)*rand(Np,N)+xmin; X=X0; %%%%%%%%%%变异操作 X1new=zeros(Np,N);%初始化 X1_new=zeros(Np,N);%初始化 X1=zeros(Np,N);%初始化 value=zeros(1,Np); while G<=Gm for i=1:Np %产生j,k,p三个不同的数 a=1; b=Np; dx=randperm(b-a+1)+a-1;%randperm()随机打乱一个数字序列 j=dx(1); k=dx(2); p=dx(3); if j==i j=dx(4); elseif k==i k=dx(4); elseif p==i p=dx(4); end %%%%%%%%%%%%%%%变异操作 namd=exp(1-Gm/(Gm+1-G));%变异算子 F=F0*2.^namd; bon=X(p,:)+F*(X(j,:)-X(k,:)); if (bon>xmin)&(bon<xmax) %防止变异超出边界 X1new(i,:)=bon; else X1new(i,:)=(xmax-xmin)*rand(1,N)+xmin; end %%%%%%%%%%%%%%%交叉操作 if rand>CR %利用二项分布来交叉 X1_new(i,:)=X(i,:); else X1_new(i,:)=X1new(i,:); end %%%%%%%%%%%%%%%选择操作 if fun(X1_new(i,:),S1,S2,S3,net_optimized,Pn_train,Tn_train)<fun(X1(i,:),S1,S2,S3,net_optimized,Pn_train,Tn_train) X1(i,:)=X1_new(i,:); value(i)=fun(X1(i,:),S1,S2,S3,net_optimized,Pn_train,Tn_train); else X1(i,:)=X(i,:); end end [fmin,nmin]=min(value); ge(G)=fmin; bestx(G,:)=X1(nmin,:); G=G+1; X=X1; end [gmin,n]=min(ge); bestvalue=gmin bestsolution=bestx(n,:) %% 解码最优个体 x = bestsolution; % 前S1*S2个编码为W1 temp1 = x(1:S1*S2); W1 = reshape(temp1,S2,S1); % 接着的S2*S3个编码为W2 temp2 = x(S1*S2+1:S1*S2+S2*S3); W2 = reshape(temp2,S3,S2); % 接着的S2个编码为B1 temp3 = x(S1*S2+S2*S3+1:S1*S2+S2*S3+S2); B1 = reshape(temp3,S2,1); %接着的S3个编码B2 temp4 = x(S1*S2+S2*S3+S2+1:S1*S2 + S2*S3 + S2 + S3); B2 = reshape(temp4,S3,1); % 设置网络初始权值和阈值 net_optimized.IW{1,1} = W1; net_optimized.LW{2,1} = W2; net_optimized.b{1} = B1; net_optimized.b{2} = B2; % 设置训练参数 net_optimized.trainParam.epochs = 3000; net_optimized.trainParam.show = 100; net_optimized.trainParam.goal = 0.001; net_optimized.trainParam.lr = 0.1; % 利用新的权值和阈值进行训练 net_optimized = train(net_optimized,Pn_train,Tn_train); %% 仿真测试 Tn_sim_optimized = sim(net_optimized,Pn_test); % 结果对比 result_optimized = [Tn_test' Tn_sim_optimized']; %均方误差 E_optimized = mse(Tn_sim_optimized - Tn_test) MAPE_optimized = mean(abs(Tn_sim_optimized-Tn_test)./Tn_sim_optimized)*100 % figure(1) % % plot(T_train,P_train(1,:),'r') % hold on % plot(T_train,P_train(3,:),'y') % hold on % plot(T_train,P_train(5,:),'b') % hold on % grid on % xlabel('标准设备的约定真值（10KP)'); % ylabel('压力传感器的输出（mv）'); % title('压力传感器的工作曲线'); % legend('t=22','t=44','t=70'); figure(2) plot(Tn_train(1:6),Pn_train(1,1:6),'r') hold on plot(Tn_train(7:12),Pn_train(1,7:12),'y') hold on plot(Tn_train(13:18),Pn_train(1,13:18),'b') hold on grid on xlabel('设备约定真值（10KP)'); ylabel('压力传感器的输出（mv）'); title('归一化后的训练样本压力传感器的工作曲线'); legend('t=22','t=44','t=70'); figure(3) plot(Tn_test(1:6),Pn_test(1,1:6),'r') hold on plot(Tn_test(7:12),Pn_test(1,7:12),'y') hold on plot(Tn_test(13:18),Pn_test(1,13:18),'b') hold on grid on xlabel('设备约定真值（10KP)'); ylabel('压力传感器的输出（mv）'); title('归一化后的测试样本压力传感器的工作曲线'); legend('t=22','t=44','t=70'); figure(4) plot(Tn_test(1:6),Tn_sim_optimized(1:6),'r')%输出DE-BP仿真结果的曲线 hold on plot(Tn_test(7:12),Tn_sim_optimized(7:12),'y') hold on plot(Tn_test(13:18),Tn_sim_optimized(13:18),'b') hold on xlabel('约定真值（10KP)'); ylabel('压力传感器的输出（mv）'); title('DE-BP的压力传感器的工作曲线'); legend('t=22','t=44','t=70'); grid on %% 未优化的BP神经网络 %net = newff(Pn_train,Tn_train,S2); net=newff(minmax(Pn_train),[6,1],{'logsig','purelin'},'traingdm');%隐含层神经元S型正切,输出层S型对数，动量梯度下降法训练BP网络, % 设置训练参数 net.trainParam.epochs = 3000; net.trainParam.show = 100; net.trainParam.goal = 0.001; net.trainParam.lr = 0.1; net=init(net); inputWeights=net.IW{1,1};% 当前输入层权值和阈值 inputbias=net.b{1}; layerWeights=net.LW{2,1};% 当前网络层权值和阈值 layerbias=net.b{2} % 利用新的权值和阈值进行训练 net = train(net,Pn_train,Tn_train); %% 仿真测试 Tn_sim = sim(net,Pn_test); %% 结果对比 result = [Tn_test' Tn_sim']; % 均方误差 E1 = mse(Tn_sim - Tn_test) MAPE1= mean(abs(Tn_sim-Tn_test)./Tn_sim)*100 % end % figure(4) % plot(T_train,P_train(1,:),'r') % hold on % plot(T_train,P_train(3,:),'y') % hold on % plot(T_train,P_train(5,:),'b') % hold on % grid on % xlabel('标准设备的约定真值（10KP)'); % ylabel('压力传感器的输出（mv）'); % title('压力传感器的工作曲线'); % legend('t=22','t=44','t=70'); figure(5) plot(Tn_test(1:6),Tn_sim(1:6),'r')%输出BP仿真结果的曲线 hold on plot(Tn_test(7:12),Tn_sim(7:12),'y') hold on plot(Tn_test(13:18),Tn_sim(13:18),'b') hold on xlabel('约定真值（10KP)'); ylabel('压力传感器的输出（mv）'); title('BP的压力传感器的工作曲线'); legend('t=22','t=44','t=70'); grid on