典型JSP车间调度问题学习10

变异操作的 Matlab 编程实现

 

三种变异操作方式：

互换

逆序：

    涉及函数 fliplr

插入：

     涉及函数： circshift

%变异操作--插入方法
function outChromes=muteChromesInsert(chromes,muteRate) 
 [pop,len]=size(chromes); 
 for i=1:pop 
 nowChrome=chromes(i,:); 
 if rand()<muteRate %判断能否变异
% outPoint=randperm(len,1);
% inPoint=randperm(len,1);
 i 
 points=randperm(len,2) 
 nowChrome 
 if points(1)<points(2) 
 insertChrome=nowChrome(points(1):points(2)-1); 
 newChrome=circshift(insertChrome,-1); 
 chromes(i,points(1):points(2)-1)=newChrome; 
 else
 insertChrome=nowChrome(points(2)-1:points(1)); 
 newChrome=circshift(insertChrome,1); 
 chromes(i,points(2)-1:points(1))=newChrome; 
 end
 chromes(i,:) 
 
 end
 outChromes=chromes; 
 end
 
end

%变异操作--逆序方法
function outChromes=muteChromesInv(chromes,muteRate) 
 [pop,len]=size(chromes); 
 for i=1:pop 
 nowChrome=chromes(i,:); 
 if rand()<muteRate 
% outPoint=randperm(len,1);
% inPoint=randperm(len,1);
 i 
 points=sortrows(randperm(len,2)')' 
 nowChrome 
 newChrome0=nowChrome(points(1):points(2)); 
 newChrome1=fliplr(newChrome0) 
 chromes(i,points(1):points(2))=newChrome1; 
 chromes(i,:) 
 
 end
 outChromes=chromes; 
 end
 
end

JSP 遗传算法集成

%找出当前种群中最优解的染色体编码以及最优解的值
function [nowOptFit,nowOptChrome]=findOptSolution(inChromes,data) 
 %step1:计算每条染色体的适应度值，放到数组中
 pop=size(inChromes,1); 
 fitnesses=zeros(pop,1); 
 for i=1:pop 
 sch=createSchedule(data,inChromes(i,:)); 
 fit=fitness(sch); 
 fitnesses(i)=fit; 
 end
 %对适应度值进行排序
 [fit2,chromeId]=sortrows(fitnesses,1); 
 nowOptFit=fit2(1); %对应最好的适应度值
 nowOptChrome=inChromes(chromeId(1),:); %对应最好的染色体
end