nsga2-带约束条件的多目标优化

logic

code

clc
clear
close all
%% 定义自变量范围
nvar=5;
nobj=2;
npop=20;
maxit=50;
pc=0.8;
nc=round(pc*npop/2)*2;
mu=0.05;
% 自变量约束条件
varmin=[1 0 3 0.6 6];
varmax=[13 2.8 21 1.6 41];
% 自变量变化步长
step=[1 0.1 0.5 0.05 1];
len=(varmax-varmin).*step;
var=[varmin;step;varmax;round(len,0)];
%% 定义结果存放模板
empty.position=[];
empty.cost=[];
empty.rank=[];
empty.domination=[];
empty.dominated=0;
empty.crowdingdistance=[];
pop=repmat(empty,npop,1);
%% 初始化种群
for i=1:npop
    pop(i).position=create_x(var);
    pop(i).cost=costfunction(pop(i).position);
end
%% 非支配排序
[pop,F]=nondominatedsort(pop);
%% 拥挤度计算
pop=calcrowdingdistance(pop,F);
%% 主程序
for it=1:maxit
    popc=repmat(empty,nc/2,2);
    
    for j=1:nc/2 
        p1=tournamentsel(pop);
        p2=tournamentsel(pop);
        [pop(j,1).position,popc(j,2).position]=crossover(p1.position,p2.position);
    end
    
    popc=popc(:);
    
    for k=1:nc
        pop(k).position=mutate(popc(k).posotion,mu,var);
        popc(k).cost=costfunction(popc(k).position);
    end
    
    newpop=[pop;popc];
    
    [pop,F]=nondominatedsort(newpop);
    pop=calcrowdingdistance(pop,F);
    
    %排序
    pop=Sortpop(pop);
    
    %淘汰
    pop=pop(1:npop);
    [pop,F]=nondominatedsort(pop);
    pop=calcrowdingdistance(pop,F);
    pop=Sortpop(pop);
    
    %更新第一等级
    F1=pop(F(1));
    
    %显示迭代信息
    disp(['Iteration' num2str(it) ':Number of F1 Members=' num2str(numel(F1))])
    
    %绘图
    figure(1)
    plotcosts(F1);
    pause(0.01);
end