黑鹰优化算法(Black eagle optimizer, BEO)

5.81KB

20 需要积分: 1

立即下载

资源介绍:

黑鹰优化算法(Black eagle optimizer, BEO)是2024年提出的一种新型的元启发式算法（智能优化算法），结合了黑鹰的生物规律和数学变换来指导粒子的搜索行为。本算法程序采用CEC2005的23个基准函数进行测试，记录多次运行的最优值、最差值、平均值和标准差等。并打印每次迭代结果。可画出函数空间分布图和算法迭代适应度收敛图。～代码完整，注释清晰，点击即可运行，可用于论文算法对比，也适用于新手学习。有问题可私聊我，看到后会回复！

function [Group_Best_Pos,Group_Best_Score,WAR_curve]=BEO(N,Max_iteration,lb,ub,dim,fobj) %% parameter setting ub = ub.*ones(dim,1); lb = lb.*ones(dim,1); R=N; T=Max_iteration; %% initialization WAR_curve=zeros(1,R); x=initialization(R,dim,ub,lb); for i=1:R Fitness1(i)=fobj(x(:,i)); end [F_x,index]=sort(Fitness1); A_Pos=x(:,index(1)); A_Score=F_x(1); t=1; L=0; A=A_Score; IT=0; %% Start iterative updates while t=ub; lb_flag=xnew(:,i)<=lb; xnew(:,i)=xnew(:,i).*(~(ub_flag+lb_flag))+(2*ub-xnew(:,i)).*ub_flag+(2*lb-xnew(:,i)).*lb_flag; end for i=1:3*R f(i)=fobj(xnew(:,i)); end [fx,indexf]=sort(f); A_Pos1=xnew(:,indexf(1)); A_Score1=fx(1); if A_Score1=0.8*ub(i)||A_Pos(i)<=0.8*lb(i) center = (ub + lb) / 2; distance = norm(A_Pos(i) - center(i)); max_distance=norm(ub(i)); min_distance=0.8*norm(ub(i)); distance_normalized=(distance-min_distance)/(max_distance-min_distance); lambda=distance_normalized*(5-15)+15; distances = sqrt((x(i,:) - A_Pos(i)).^2); [~, sorted_indices1] = sort(distances); xnew3(i,:) = zeros(1,R); poissrn_values = poissrnd(lambda, 1, R); for j = 1:R xnew3(i,j) = A_Pos(i) + (x(i,sorted_indices1(j)) - A_Pos(i)) .* (1 + poissrn_values(j)); end lambda_normalized=(lambda-5)/(15-5); bate=sin(lambda_normalized*pi/2)*(0.953-0.89)+0.89; xnew3(i,:)=xnew3(i,:)+bate*(A_Pos(i)-xnew3(i,:)); x0(i,:)=xnew3(i,:); end end %% hovering for j=1:R rotation_angle = rand*2*pi; rotation_matrix = eye(dim); for ri = 1:dim-1 rotation_matrix(ri,ri) = cos(rotation_angle); rotation_matrix(ri,ri+1) = -sin(rotation_angle); rotation_matrix(ri+1,ri) = sin(rotation_angle); rotation_matrix(ri+1,ri+1) = cos(rotation_angle); end xnew4(:,j) = A_Pos +rotation_matrix * (x(:,j) - A_Pos); end for i=1:R ub_flag=xnew4(:,i)>=ub; lb_flag=xnew4(:,i)<=lb; xnew4(:,i)=xnew4(:,i).*(~(ub_flag+lb_flag))+(2*ub-xnew4(:,i)).*ub_flag+(2*lb-xnew4(:,i)).*lb_flag; f=fobj(xnew4(:,i)); if f0; xnew(i,:)=xnew5(i,:).*(~flag1)+x0(i,:); end for i=1:R ub_flag=xnew(:,i)>=ub; lb_flag=xnew(:,i)<=lb; xnew(:,i)=xnew(:,i).*(~(ub_flag+lb_flag))+(lb+rand(dim,1).*(ub-lb)).*ub_flag+(lb+rand(dim,1).*(ub-lb)).*lb_flag; f=fobj(xnew(:,i)); if fub; lb_flag=xnew6(:,i)=0.1*T||L>0.1*T/2 for j=1:R Fitness(j)=fobj(x(:,j)); z=1/(2*exp(-(A_Score/(Fitness(j)+eps)))); s=-ones(dim,1); s0=s+2*rand(dim,1); xnew7(:,j)=A_Pos+z.*s0.*(x(:,j)-(0.4+0.6.*tent_map(rand(dim,1), 0.5)).*A_Pos); end for i=1:R ub_flag=xnew7(:,i)>ub; lb_flag=xnew7(:,i)=0.3*T||L>0.3*T/2 f = 1./(1*(1+exp(-(14*t/T-9))))+0.3;%0.3-1.3 for j=1:R if mod(j,2)==1 r5=tent_map(rand(1), 0.5); xnew8(:,j)=x(:,j)+r5*f*(A_Pos-x(:,j))+rand(dim,1).*cos((pi*j)/R).*(A_Pos-x(:,j)); else r6=tent_map(rand(1), 0.5); xnew8(:,j)=x(:,j)+r6*f*(A_Pos-x(:,j))+randn(dim,1).*sin((pi*j)/R).*(A_Pos-x(:,j)); end end for i=1:R ub_flag=xnew8(:,i)>ub; lb_flag=xnew8(:,i)ub; lb_flag=xnew9(:,i)