大连海事大学作业汇编-遗传算法源代码.docxVIP

一. 遗传算法源代码1. 主程序：clcclearall;closeall;popsize=60;%种群规模chromlength=24;%染色体长度J=4; %备选配送中心数P=2; %预计建设数量I=6; %顾客数量B=1000000; %投资预算a=[400 300 500 300];c1=[100 80 90 120];c2=[120 110 130 140 150 140 150 160 120 110 150 130 160 120 110 170 120 180 130 130 130 130 130 130]; d=[200 230 180 150 300 260 200 230 180 150 300 260 200 230 180 150 300 260 200 230 180 150 300 260];e=[10000 11000 12000 11000];pm=0.3;%变异概率pc=0.5;%交叉概率[pop0,initpop]=init(popsize,chromlength,J,I); %随机产生初始种群 for i=1:300 %300为遗传代数[fit,cumsump]=fitness(initpop,pop0,J,P,I,B,a,c1,c2,d,e);%计算种群中个体适应度[bestpop,bestnewpop,bestfit]=best(pop0,initpop,fit);%记录最佳适应度[pop1,newpop] =selection(pop0,initpop,cumsump); %选择newpop1=crossover(pop1,newpop,pc); %交叉[pop2,newpop2]=mutation(pop1,newpop1,pm); %变异pop0=pop2;initpop=newpop2;enddisp(最小成本为：)bestfitdisp(配送中心选择方案：)bestpopdisp(顾客配送方案：)bestnewpop2. 生成初代种群子函数function [pop0,initpop]=init(popsize,chromlength,J,I);initpop=zeros(popsize,chromlength);pop=zeros(popsize,chromlength);%%%%%%%%%%生成配送中心选择方案的初代解%%%%%%%%pop0=round(rand(popsize,J));for i=1:popsizeif pop0(i,:)==[0 0 0 0]pop0(i,:)=[1 1 1 1];endend%%%%%%%%%%生成顾客配送方案的初代解%%%%%%%%for i=1:popsize for j=1:Jif pop0(i,j)==1;pop(i,((j-1)*I+1):((j-1)*I+6))=round(rand(1,I));endendendinitpop=pop; 3. 种群中个体适应度子函数function [fit,cumsump]=fitness(initpop,pop0,J,P,I,B,a,c1,c2,d,e);[px,py]=size(initpop);fit=zeros(px,1);dudu=zeros(px,I);for i=1:px p=sum(pop0(i,:)); b=sum(pop0(i,:).*e); a=0;%%%%%%%%%适应度函数%%%%%%%%%for j=1:I dudu(i,j)=dudu(i,j)+initpop(i,j)+initpop(i,6+j)+initpop(i,12+j)+initpop(i,18+j);ifdudu(i,j)==0 a=1;endendif ( bB | p1 | pP | a==1) fit(i)=500000;else fit(i)=sum(a.*c1.*pop0(i,:)+e.*pop0(i,:))+sum(initpop(i,:).*d.*c2);endend%%%%%%%%%种群选择准备工作%%%%%%fit1=600000-fit;fsum=sum(fit1);pp=fit1/fsum;cumsump=cumsum(pp);4. 记录最优解子函数function [bestpop,bestnewpop,bestfit]=best(pop0,initpop,fit); [px,py]=size(pop0); [bestfit,num