2025年3月一区SCI-梦境优化算法Dream Optimization Algorithm-附Matlab免费代码-CSDN博客

本文链接：https://blog.csdn.net/weixin_44028734/article/details/145572744

引言

本期介绍了一种新的受人类做梦行为启发的求解优化问题的元启发式算法-梦境优化算法Dream Optimization Algorithm ，DOA，梦具有部分记忆保留、遗忘和逻辑自组织等特征，与元启发式算法中的优化过程有很强的相似性。该算法于2025年3月最新发表在中科院1区SCI期刊 Computer Methods in Applied Mechanics and Engineering。

参考文献

[1] Yifan Lang, Yuelin Gao. Dream Optimization Algorithm (DOA): A novel metaheuristic optimization algorithm inspired by human dreams and its applications to real-world engineering problems. Computer Methods in Applied Mechanics and Engineering. https://doi.org/10.1016/j.cma.2024.117718.

Matlab代码下载

微信搜索并关注-优化算法侠（英文名：Swarm-Opti），或扫描下方二维码关注，以算法名字搜索历史文章即可下载。

function [fbest,sbest,fbest_history] = DOA(pop,T,lb,ub,D,fobj);
% 关注微信公众号：优化算法侠
%pop stands for population size
%T represents the maximum number of population iteration
%lb represents the lower bound of the search space
%ub represents the upper bound of the search space
%D represents the dimension of the problem
%fobj represents the objective function
lb=lb.*ones(1,D);   
ub=ub.*ones(1,D);                            %Ensure that the dimensions of the upper and lower bounds match the dimensionality of the problem
x =initialization(pop,D,ub,lb);              %Initialize the population
SELECT=1:pop;                                      
sbest=ones(1,D);                             %sbest is the current optimal solution
sbestd=ones(5,D);                            %sbestd represents the current optimal solution for each of the 5 groups
fbest=inf;                                   %fbest is the current optimal value
fbestd=ones(5,1);                            %fbestd represents the current optimal value for each of the 5 groups
fbest_history=ones(1,T);                     %fbest_history is a vector composed of the best values of each generation
for i = 1:5
      fbestd(i)=fbest;
end   
              
for i=1:(9*T/10)                                               %Exploration phase
      for m=1:5                                                %Divide into 5 groups       
         k=randi([ceil(D/8/m),ceil(D/3/m)]);            
         for j=(((m-1)/5*pop)+1) : (m/5*pop)                 
                     if (fobj(x(j,:)) <fbestd(m))
                      sbestd(m,:)=x(j,:);                         
                      fbestd(m)=fobj(x(j,:));                  %Update the optimal value and solution for each group
                     end
         end
          for j=(((m-1)/5*pop)+1) : (m/5*pop)
           x(j,:)=sbestd(m,:);                                  %Memory strategy
           in=randperm(D,k);        
           if rand<0.9
               for h=1:k
                   x(j,in(h))=x(j,in(h))+(rand*(ub(in(h))-lb(in(h)))+lb(in(h)))*(cos((1*i+T/10)*pi/T)+1)/2;      %Forgetting and supplementation strategy
                if (x(j,in(h))>ub(in(h))) | (x(j,in(h))<lb(in(h)))   
                    if D>15                                      %The boundary handling method when the problem dimensionality>15
                     select=SELECT;     
                     select(j)=[];
                     sel=select(randi(pop-1));                             
                     x(j,in(h))=x(sel,in(h)); 
                    else                                          %The boundary handling method when the problem dimensionality<=15
                     x(j,in(h))=rand*(ub(in(h))-lb(in(h)))+lb(in(h));                
                    end  
                 end
               end
            else
               for h=1:k                       
                   x(j,in(h))=x(randi(pop),in(h));                                      
               end
             end
          end     
            if (fbestd(m)<fbest)            %Update the optimal value and solution for the entire population
               fbest=fbestd(m);
               sbest=sbestd(m,:);
            end
      end                                               
fbest_history(i)=fbest;                                                                  
end
for i=((9*T/10)+1):T                        %Exploitation phase
      for p=1 : pop                     
          if (fobj(x(p,:)) <fbest)          %Update the optimal value and solution for the entire population
          sbest=x(p,:);
          fbest=fobj(x(p,:));
          end                                                                                             
      end                               
    for j=1:pop
        fitness(j,:)=fobj(x(j,:));
        km=max(2,ceil(D/3));
        k=randi([2,km]);
        x(j,:)=sbest;
        in=randperm(D,k); 
        for h=1:k
            x(j,in(h))=x(j,in(h))+(rand*(ub(in(h))-lb(in(h)))+lb(in(h)))*(cos((i)*pi/T)+1)/2;     %Forgetting and supplementation strategy              
            if (x(j,in(h))>ub(in(h))) | (x(j,in(h))<lb(in(h)))   
                if D>15                                      %The boundary handling method when the problem dimensionality>15
                select=SELECT;     
                select(j)=[];
                sel=select(randi(pop-1));                             
                x(j,in(h))=x(sel,in(h)); 
                else                                          %The boundary handling method when the problem dimensionality<=15
                   x(j,in(h))=rand*(ub(in(h))-lb(in(h)))+lb(in(h));                
                end  
            end
        end
      end
fbest_history(i)=fbest;                                                                  
end          
end

2025年3月一区SCI-梦境优化算法Dream Optimization Algorithm-附Matlab免费代码