基于matlab平台的启发式智能算法_蝙蝠算法程序_matlab

% ========ywd 整理自新兴元启发式优化算法=========== % % Files of the Matlab programs included in the book: % % Xin-She Yang, Nature-Inspired Metaheuristic Algorithms, % % Second Edition, Luniver Press, (2010). www.luniver.com % % ======================================================== % % -------------------------------------------------------- % % Bat-inspired algorithm for continuous optimization (demo)% % Programmed by Xin-She Yang @Cambridge University 2010 % % -------------------------------------------------------- % % Usage: bat_algorithm([20 0.25 0.5]); % % Display help % help bat_algorithm.m clc clear %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); % 参数初始化 para=([25,0.25,0.45]); n=para(1); % Population size, typically 10 to 25 种群数量，一般取10-25 A=para(2); % Loudness (constant or decreasing) 音量（常数或者递减） r=para(3); % Pulse rate (constant or decreasing) 脉冲率（常数或者递减） % This frequency range determines the scalings 确定频率变化范围 Qmin=0; % Frequency minimum 频率最小值 Qmax=2; % Frequency maximum 频率最大值 % Iteration parameters 算法迭代参数设置 tol=0.01; % Stop tolerance 算法结束误差 N_iter=0; % Total number of function evaluations 算法运行总次数 % Dimension of the search variables 变量的参数 d=21; % Initial arrays 数组初始化 Q=zeros(n,1); % Frequency 频率值 v=zeros(n,d); % Velocities 速度值 for i=1:n sol(i,:)=randn(1,21); %sol(i,:)=5*rands(1,21);%更改rands前的参数 v(i,:)=rands(1,21); Q(i,:)=rands(1,1); fitness(i)=fun(sol(i,:),inputnum,hiddennum,outputnum,net,inputn,outputn); end [fmin,I]=min(fitness); best=sol(I,:); maxgen=500; % ====================================================== % % Note: As this is a demo, here we did not implement the % % reduction of loudness and increase of emission rates. % % Interested readers can do some parametric studies % % and also implementation various changes of A and r etc % % ====================================================== % % Start the iterations -- Bat Algorithm 开始迭代 while (fmin>tol & N_iter<1000) for j=1:maxgen % Loop over all bats/solutions for i=1:n, Q(i)=Qmin+(Qmin-Qmax)*rand; v(i,:)=v(i,:)+(sol(i,:)-best)*Q(i); S(i,:)=sol(i,:)+v(i,:); % Pulse rate if rand>r S(i,:)=best+0.01*randn(1,d); end % Evaluate new solutions 评估新解的适应度值 Fnew=fun(S(i,:),inputnum,hiddennum,outputnum,net,inputn,outputn); % If the solution improves or not too loudness if (Fnew<=fitness(i)) & (rand<A) , sol(i,:)=S(i,:); fitness(i)=Fnew; end % Update the current best 更新当前最优解 if Fnew<=fmin, best=S(i,:); fmin=Fnew; end yy(i)=fmin; end N_iter=N_iter+n; end end % Output/display 输出显示结果 disp(['Number of evaluations: ',num2str(N_iter)]); disp(['Best =',num2str(best),' fmin=',num2str(fmin)]); x=best; 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; net=train(net,inputn,outputn); %% BP网络预测 %预测数据归一化 inputn_test=mapminmax('apply',input_test,inputps); %网络预测输出 an=sim(net,inputn_test); %网络输出反归一化 BPoutput=mapminmax('reverse',an,outputps); AE=mean(output_test-BPoutput); MAE=mean(abs(output_test-BPoutput)); MSE=mean((output_test-BPoutput).^2); MAPE=mean(abs((output_test-BPoutput)/output_test)); %%%%% ============ end ====================================