1 简介
工业过程常含有显著的非线性,时变等复杂特性,传统的核极限学习机有时无法充分利用数据信息,所建软测量模型预测性能较差.为了提高核极限学习机的泛化能力和预测精度,提出一种哈里斯鹰算法结合核极限学习机软测量建模方法.通过哈里斯鹰优化极限学习机的惩罚系数和核宽,得到一组最优超参数;最后将该方法应用于脱丁烷塔过程软测量建模中.仿真结果表明,优化后的核极限学习机模型预测精度有明显的提高,验证了所提方法不仅是可行的,而且具有良好的预测精度和泛化性能.
2 部分代码
%%function[Rabbit_Energy,Rabbit_Location,CNVG] =NCHHO_IoV(N,T,lb,ub,dim,fobj)% initialize the location and Energy of the rabbitRabbit_Location=zeros(1,dim);Rabbit_Energy=0;%Initialize the locations of Harris' hawksX=initialization(N,dim,ub,lb);CNVG=zeros(1,T);t=0;% Loop counterwhilet<Tfori=1:size(X,1)% Check boundriesFU=X(i,:)>ub;FL=X(i,:)<lb;X(i,:)=(X(i,:).*(~(FU+FL)))+ub.*FU+lb.*FL;% fitness of locationsfitness=fobj(X(i,:));% Update the location of Rabbitiffitness>Rabbit_EnergyRabbit_Energy=fitness;Rabbit_Location=X(i,:);endendE1=abs(2*(1-(t/T))-2);% factor to show the decreaing energy of rabbita1=4; % Initial chaotic map parameter configurationteta=0.7; % Initial chaotic map parameter configuration% Update the location of Harris' hawksfori=1:size(X,1)forii=1:4Cm(1,ii) =abs((a1/4)*sin(pi*teta));teta=Cm(1,ii);endE0=2*rand()-1;%-1<E0<1Escaping_Energy=E1*(E0); % escaping energy of rabbitifabs(Escaping_Energy)>=1%% Exploration:% Harris' hawks perch randomly based on 2 strategy:q=rand();rand_Hawk_index=floor(N*rand()+1);X_rand=X(rand_Hawk_index, :);ifq<0.5% perch based on other family membersX(i,:)=X_rand-Cm(1,1)*abs(X_rand-2*Cm(1,2)*X(i,:));elseifq>=0.5% perch on a random tall tree (random site inside group's home range)X(i,:)=(Rabbit_Location(1,:)-mean(X))-Cm(1,3)*((ub-lb)*Cm(1,4)+lb);endelseifabs(Escaping_Energy)<1%% Exploitation:% Attacking the rabbit using 4 strategies regarding the behavior of the rabbit%% phase 1: surprise pounce (seven kills)% surprise pounce (seven kills): multiple, short rapid dives by different hawksr=rand();% probablity of each eventifr>=0.5&&abs(Escaping_Energy)<0.5% Hard besiegeX(i,:)=(Rabbit_Location)-Escaping_Energy*abs(Rabbit_Location-X(i,:));endifr>=0.5&&abs(Escaping_Energy)>=0.5% Soft besiegeJump_strength=2*(1-rand());% random jump strength of the rabbitX(i,:)=(Rabbit_Location-X(i,:))-Escaping_Energy*abs(Jump_strength*Rabbit_Location-X(i,:));end%% phase 2: performing team rapid dives (leapfrog movements)ifr<0.5&&abs(Escaping_Energy)>=0.5% Soft besiege % rabbit try to escape by many zigzag deceptive motionsw1=2*exp(-(8*t/T)^2); % Non-linear control ParameterJump_strength=2*(1-rand());X1=w1*Rabbit_Location-Escaping_Energy*abs(Jump_strength*Rabbit_Location-X(i,:));iffobj(X1)>fobj(X(i,:))% improved move?X(i,:)=X1;else% hawks perform levy-based short rapid dives around the rabbitX2=w1*Rabbit_Location-Escaping_Energy*abs(Jump_strength*Rabbit_Location-X(i,:))+rand(1,dim).*Levy(dim);if(fobj(X2)>fobj(X(i,:)))% improved move?X(i,:)=X2;endendendifr<0.5&&abs(Escaping_Energy)<0.5% Hard besiege % rabbit try to escape by many zigzag deceptive motions% hawks try to decrease their average location with the rabbitw1=2*exp(-(8*t/T)^2);Jump_strength=2*(1-rand());X1=w1*Rabbit_Location-Escaping_Energy*abs(Jump_strength*Rabbit_Location-mean(X));iffobj(X1)>fobj(X(i,:))% improved move?X(i,:)=X1;else% Perform levy-based short rapid dives around the rabbitX2=w1*Rabbit_Location-Escaping_Energy*abs(Jump_strength*Rabbit_Location-mean(X))+rand(1,dim).*Levy(dim);if(fobj(X2)>fobj(X(i,:)))% improved move?X(i,:)=X2;endendend%%endendt=t+1;CNVG(t)=Rabbit_Energy;endend% ___________________________________functiono=Levy(d)beta=1.5;sigma=(gamma(1+beta)*sin(pi*beta/2)/(gamma((1+beta)/2)*beta*2^((beta-1)/2)))^(1/beta);u=randn(1,d)*sigma;v=randn(1,d);step=u./abs(v).^(1/beta);o=step;end
3 仿真结果
4 参考文献
[1]彭甜等. "基于改进哈里斯鹰算法优化ELM的风速预测方法及系统.".
部分理论引用网络文献，若有侵权联系博主删除。
5 完整MATLAB代码与数据下载地址
见博客主页头条
