混合粒子群算法+粒子群参数设置改进策略（matlab代码实现）

function [fitnessgbest,gen,record] = test_Original( maxgen,popmax,popmin ,index,esp) %UNTITLED2 此处显示有关此函数的摘要 % 此处显示详细说明 %% 原算法 %tic; gen=maxgen; %esp=10^-15; Dim = 2; %维度（自变量的个数） %index=1; %测试函数索引（对应表里面的函数） wmax=0.9; wmin=0.4; cmax=1.5; cmin=0.5; %maxgen = 50; %迭代次数50 sizepop = 100; %种群规模100 Vmax = 20; %粒子移动最大速度5 Vmin = -20; %粒子移动最小速度-5 %popmax = 100; %自变量最大取值（搜索范围） %popmin = 50; %自变量最小取值 pc=0.7; pm=0.3; N=pc*sizepop; M=pm*sizepop; record=zeros(1,maxgen); %% 产生初始粒子和速度 for i = 1:sizepop % 随机产生一个种群 pop(i,:) = (popmax-popmin)*rand(1,Dim)+popmin; %初始种群 V(i,:) = (Vmax-Vmin)*rand(1,Dim)+Vmin; %初始化速度 % 计算适应度 fitness(i) = fun(pop(i,:),index); %计算适应度(这里设定的是结果的值越小=适应度越低) end %% 个体极值和群体极值 [bestfitness bestindex] = min(fitness); %bestindex:全局最优粒子索引 gbest = pop(bestindex,:); %全局最佳位置 pbest = pop; %个体最佳 fitnesspbest = fitness; %个体最佳适应度值 fitnessgbest = bestfitness; %全局最佳适应度值 %% 迭代寻优 for i = 1:maxgen %代数更迭 for j = 1:sizepop %遍历个体 % 速度更新 w=wmax-(wmax-wmin)*i/maxgen; c1=cmax-(cmax-cmin)*i/maxgen; c2=cmax-(cmax-cmin)*i/maxgen; V(j,:) = w*V(j,:) + c1*rand*(pbest(j,:) - pop(j,:)) + c2*rand*(gbest - pop(j,:)); %速度边界处理 V(j,find(V(j,:)>Vmax)) = Vmax; V(j,find(V(j,:)<Vmin)) = Vmin; % 种群更新 pop(j,:) = pop(j,:) + V(j,:); %位置边界处理 pop(j,find(pop(j,:)>popmax)) = popmax; pop(j,find(pop(j,:)<popmin)) = popmin; % 适应度值更新 fitness(j) = fun(pop(j,:),index);%计算函数值 end for j = 1:sizepop % 个体最优更新 if fitness(j) < fitnesspbest(j) pbest(j,:) = pop(j,:); fitnesspbest(j) = fitness(j); end % 群体最优更新 if fitness(j) < fitnessgbest gbest = pop(j,:); fitnessgbest = fitness(j); for j=1:sizepop GApop(j,:)=pop(j,:); [n m]=size(GApop); end c=randperm(sizepop); b=c(1:N); e=b'; for z=1:N GApop1(z,:)=GApop(e(z),:); end [n1 m1]=size(GApop1); thebest=gbest'*ones(1,n1); GApops=0.5*GApop1+0.5*thebest'; %进行变异操作 % max1=max(max(gbest)); % min1=min(min(gbest)); % max2=round(max1); % min2=round(min1); c3=randperm(sizepop); b1=c3(1:M); e1=b1'; for z=1:M GApop2(z,:)=GApop(e1(z),:); end GApops1=GApop2+randi([popmin,popmax],M,Dim); GApops2=[GApops;GApops1]; pop=GApops2; end record(1,i)=fitnessgbest; %采取1/fitnessgbest是因为我们设定了值越小，适应度越大。 if fitnessgbest<=esp gen=i; esp=-inf; end end %toc; end % % end