MOEA.zip_algorithm_matlab_moea_多目标_多目标进化算法

% % input Pt ,size of N*m;output Rt,size of 2N*m function [ Rt ] = Genetic_Operators( Pt,minx,maxx,numvariate,numfun) %% 不交叉则变异，避免种群中数值相同的个体存在，这样可以保持种群的多样性 pc=0.9; pm=0.1; cetac=20; cetam=20; maxminx=maxx-minx; N=size(Pt,1); numvariate; Qt=Pt; crossPt=zeros(N,numvariate); for i=1:numvariate crossPt(:,i)=randperm(N); end detamax=10;%这个参数有待调整 for j=1:numvariate %%先交叉 for i=1:ceil(N/2) pccross=rand(1); if(pccross<pc) ui=rand(1); if ui<=0.5 betau=(2*ui)^(1/(cetac+1)); else if ui==1 ui=1-0.005; end betau=(1/(2*(1-ui)))^(1/(cetac+1)); end Qt1=0.5*((1-betau)*Pt(crossPt(i,j),j)+(1+betau)*Pt(N-crossPt(i,j)+1,j)); Qt2=0.5*((1+betau)*Pt(crossPt(i,j),j)+(1-betau)*Pt(N-crossPt(i,j)+1,j)); Qt(2*i-1,j)=mod(Qt1-minx,maxminx)+minx; Qt(2*i,j)=mod(Qt2-minx,maxminx)+minx; else pmu1=rand(1); if pmu1<0.5 deta1=(2*pmu1)^(1/(cetam+1))-1; else deta1=1-(2*(1-pmu1))^(cetam+1); end pmu2=rand(1); if pmu2<0.5 deta2=(2*pmu2)^(1/(cetam+1))-1; else deta2=1-(2*(1-pmu2))^(cetam+1); end Qt1=Pt(crossPt(i,j),j)+deta1*detamax; Qt2=Pt(N-crossPt(i,j)+1,j)+deta2*detamax; Qt(2*i-1,j)=mod(Qt1-minx,maxminx)+minx; Qt(2*i,j)=mod(Qt2-minx,maxminx)+minx; end end end funvalueQt=funfvalue(Qt(:,1:numvariate)); Qt(:,(numvariate+1):(numvariate+numfun))=funvalueQt; Rt=cat(1,Pt,Qt); end