遗传算法的经典案例Word文档下载推荐.docx
《遗传算法的经典案例Word文档下载推荐.docx》由会员分享,可在线阅读,更多相关《遗传算法的经典案例Word文档下载推荐.docx(16页珍藏版)》请在冰豆网上搜索。
else
farm_X(j+1,i)=1;
end
end
8
ifRAND<
1/3
farm_Y(j,3*i-2)=1;
elseifRAND>
2/3
farm_Y(j,3*i)=1;
farm_Y(j,3*i-1)=1;
counter=0;
%设置迭代计数器
whilecounter<
M%停止条件为达到最大迭代次数
%第三步:
交叉
newfarm_X=zeros(14,N);
newfarm_Y=zeros(8,3*N);
Ser=randperm(N);
%对X做交叉
fori=1:
(N-1)
A_X=farm_X(:
Ser(i));
B_X=farm_X(:
Ser(i+1));
cp=2*unidrnd(6);
a_X=[A_X(1:
cp);
B_X((cp+1):
end)];
b_X=[B_X(1:
A_X((cp+1):
newfarm_X(:
i)=a_X;
i+1)=b_X;
%对Y做交叉
A_Y=farm_Y(:
(3*Ser(i)-2):
(3*Ser(i)));
B_Y=farm_Y(:
(3*Ser(i+1)-2):
(3*Ser(i+1)));
cp=unidrnd(7);
a_Y=[A_Y(1:
cp,:
);
B_Y((cp+1):
end,:
)];
b_Y=[B_Y(1:
A_Y((cp+1):
newfarm_Y(:
(3*i-2):
(3*i))=a_Y;
(3*i+1):
(3*i+3))=b_Y;
%新旧种群合并
FARM_X=[farm_X,newfarm_X];
FARM_Y=[farm_Y,newfarm_Y];
%第四步:
选择复制
Ser=randperm(2*N);
FITNESS=zeros(1,2*N);
fitness=zeros(1,N);
(2*N)
X=FARM_X(:
i);
Y=FARM_Y(:
(3*i));
FITNESS(i)=COST(X,Y,x1_x14,F_x1_x14,A,Q,C,S,b);
f1=FITNESS(Ser(2*i-1));
f2=FITNESS(Ser(2*i));
iff1<
f2
farm_X(:
i)=FARM_X(:
Ser(2*i-1));
farm_Y(:
(3*i))=FARM_Y(:
(3*Ser(2*i-1)-2):
(3*Ser(2*i-1)));
fitness(i)=f1;
Ser(2*i));
(3*Ser(2*i)-2):
(3*Ser(2*i)));
fitness(i)=f2;
%记录最佳个体和收敛曲线
minfitness=min(fitness);
meanfitness=mean(fitness);
LC1(counter+1)=minfitness;
LC2(counter+1)=meanfitness;
pos=find(fitness==minfitness);
Xp=farm_X(:
pos
(1));
Yp=farm_Y(:
(3*pos
(1)-2):
(3*pos
(1)));
Zp=minfitness;
%第五步:
变异
ifPm>
rand
GT_X=farm_X(:
GT_Y=farm_Y(:
pos1=2*unidrnd(7);
ifGT_X(pos1)==1
GT_X(pos1-1)=1;
GT_X(pos1)=0;
i)=GT_X;
elseifGT_X(pos1)==0
GT_X(pos1-1)=0;
GT_X(pos1)=1;
pos2=unidrnd(8);
GT_Y(pos2,:
)=zeros(1,3);
GT_Y(pos2,unidrnd(3))=1;
end
counter=counter+1
Xp=Xp'
;
Yp=Yp'
%plot(LC1)
%holdon
plot(LC2)
遗传算法程序:
说明:
fga.m为遗传算法的主程序;
采用二进制Gray编码,采用基于轮盘赌法的非线性排名选择,均匀交叉,变异操作,而且还引入了倒位操作!
function[BestPop,Trace]=fga(FUN,LB,UB,eranum,popsize,pCross,pMutation,pInversion,options)
%[BestPop,Trace]=fmaxga(FUN,LB,UB,eranum,popsize,pcross,pmutation)
%Findsamaximumofafunctionofseveralvariables.
%fmaxgasolvesproblemsoftheform:
%maxF(X)subjectto:
LB<
=X<
=UB
%BestPop-最优的群体即为最优的染色体群
%Trace-最佳染色体所对应的目标函数值
%FUN-目标函数
%LB-自变量下限
%UB-自变量上限
%eranum-种群的代数,取100--1000(默认200)
%popsize-每一代种群的规模;
此可取50--200(默认100)
%pcross-交叉概率,一般取0.5--0.85之间较好(默认0.8)
%pmutation-初始变异概率,一般取0.05-0.2之间较好(默认0.1)
%pInversion-倒位概率,一般取0.05-0.3之间较好(默认0.2)
%options-1*2矩阵,options
(1)=0二进制编码(默认0),option
(1)~=0十进制编
%码,option
(2)设定求解精度(默认1e-4)
%
%------------------------------------------------------------------------
T1=clock;
ifnargin<
3,error('
FMAXGArequiresatleastthreeinputarguments'
ifnargin==3,eranum=200;
popsize=100;
pCross=0.8;
pMutation=0.1;
pInversion=0.15;
options=[01e-4];
ifnargin==4,popsize=100;
ifnargin==5,pCross=0.8;
ifnargin==6,pMutation=0.1;
ifnargin==7,pInversion=0.15;
iffind((LB-UB)>
0)
error('
数据输入错误,请重新输入(LB<
UB):
'
s=sprintf('
程序运行需要约%.4f秒钟时间,请稍等......'
(eranum*popsize/1000));
disp(s);
globalmnNewPopchildren1children2VarNum
bounds=[LB;
UB]'
bits=[];
VarNum=size(bounds,1);
precision=options
(2);
%由求解精度确定二进制编码长度
bits=ceil(log2((bounds(:
2)-bounds(:
1))'
./precision));
%由设定精度划分区间
[Pop]=InitPopGray(popsize,bits);
%初始化种群
[m,n]=size(Pop);
NewPop=zeros(m,n);
children1=zeros(1,n);
children2=zeros(1,n);
pm0=pMutation;
BestPop=zeros(eranum,n);
%分配初始解空间BestPop,Trace
Trace=zeros(eranum,length(bits)+1);
i=1;
whilei<
=eranum
m
value(j)=feval(FUN(1,:
),(b2f(Pop(j,:
),bounds,bits)));
%计算适应度
[MaxValue,Index]=max(value);
BestPop(i,:
)=Pop(Index,:
Trace(i,1)=MaxValue;
Trace(i,(2:
length(bits)+1))=b2f(BestPop(i,:
),bounds,bits);
[selectpop]=NonlinearRankSelect(FUN,Pop,bounds,bits);
%非线性排名选择
[CrossOverPop]=CrossOver(selectpop,pCross,round(unidrnd(eranum-i)/eranum));
%采用多点交叉和均匀交叉,且逐步增大均匀交叉的概率
%round(unidrnd(eranum-i)/eranum)
[MutationPop]=Mutation(CrossOverPop,pMutation,VarNum);
%变异
[InversionPop]=Inversion(MutationPop,pInversion);
%倒位
Pop=InversionPop;
%更新
pMutation=pm0+(i^4)*(pCross/3-pm0)/(eranum^4);
%随着种群向前进化,逐步增大变异率至1/2交叉率
p(i)=pMutation;
i=i+1;
t=1:
eranum;
plot(t,Trace(:
1)'
title('
函数优化的遗传算法'
xlabel('
进化世代数(eranum)'
ylabel('
每一代最优适应度(maxfitness)'
[MaxFval,I]=max(Trace(:
1));
X=Trace(I,(2:
length(bits)+1));
holdon;
plot(I,MaxFval,'
*'
text(I+5,MaxFval,['
FMAX='
num2str(MaxFval)]);
str1=sprintf('
进化到%d代,自变量为%s时,得本次求解的最优值%f\n对应染色体是:
%s'
I,num2str(X),MaxFval,num2str(BestPop(I,:
)));
disp(str1);
%figure
(2);
plot(t,p);
%绘制变异值增大过程
T2=clock;
elapsed_time=T2-T1;
ifelapsed_time(6)<
elapsed_time(6)=elapsed_time(6)+60;
elapsed_time(5)=elapsed_time(5)-1;
ifelapsed_time(5)<
elapsed_time(5)=elapsed_time(5)+60;
elapsed_time(4)=elapsed_time(4)-1;
end%像这种程序当然不考虑运行上小时啦
str2=sprintf('
程序运行耗时%d小时%d分钟%.4f秒'
elapsed_time(4),elapsed_time(5),elapsed_time(6));
disp(str2);
TSP问题遗传算法matlab源程序(注释详细,经反复实验收敛速度快)
%TSP问题(又名:
旅行商问题,货郎担问题)遗传算法通用matlab程序
%D是距离矩阵,n为种群个数,建议取为城市个数的1~2倍,
%C为停止代数,遗传到第C代时程序停止,C的具体取值视问题的规模和耗费的时间而定
%m为适应值归一化淘汰加速指数,最好取为1,2,3,4,不宜太大
%alpha为淘汰保护指数,可取为0~1之间任意小数,取1时关闭保护功能,最好取为0.8~1.0
%R为最短路径,Rlength为路径长度
function[R,Rlength]=geneticTSP(D,n,C,m,alpha)
[N,NN]=size(D);
farm=zeros(n,N);
%用于存储种群
n
farm(i,=randperm(N);
%随机生成初始种群
R=farm(1,;
%存储最优种群
len=zeros(n,1);
%存储路径长度
fitness=zeros(n,1);
%存储归一化适应值
C
len(i,1)=myLength(D,farm(i,);
%计算路径长度
maxlen=max(len);
minlen=min(len);
fitness=fit(len,m,maxlen,minlen);
%计算归一化适应值
rr=find(len==minlen);
R=farm(rr(1,1),;
%更新最短路径
FARM=farm;
%优胜劣汰,nn记录了复制的个数
nn=0;
iffitness(i,1)>
=alpha*rand
nn=nn+1;
FARM(nn,=farm(i,;
FARM=FARM(1:
nn,;
[aa,bb]=size(FARM);
%交叉和变异
whileaa<
ifnn<
=2
nnper=randperm
(2);
nnper=randperm(nn);
A=FARM(nnper
(1),;
B=FARM(nnper
(2),;
[A,B]=intercross(A,B);
FARM=[FARM;
A;
B];
ifaa>
n,;
%保持种群规模为n
farm=FARM;
clearFARM
Rlength=myLength(D,R);
function[a,b]=intercross(a,b)
L=length(a);
ifL<
=10%确定交叉宽度
W=1;
elseif((L/10)-floor(L/10))>
=rand&
&
L>
10
W=ceil(L/10);
else
W=floor(L/10);
p=unidrnd(L-W+1);
%随机选择交叉范围,从p到p+W
W%交叉
x=find(a==b(1,p+i-1));
y=find(b==a(1,p+i-1));
[a(1,p+i-1),b(1,p+i-1)]=exchange(a(1,p+i-1),b(1,p+i-1));
[a(1,x),b(1,y)]=exchange(a(1,x),b(1,y));
function[x,y]=exchange(x,y)
temp=x;
x=y;
y=temp;
%计算路径的子程序
functionlen=myLength(D,p)
len=D(p(1,N),p(1,1));
fori=1N-1)
len=len+D(p(1,i),p(1,i+1));
%计算归一化适应值子程序
functionfitness=fit(len,m,maxlen,minlen)
fitness=len;
length(len)
fitness(i,1)=(1-((len(i,1)-minlen)/(maxlen-minlen+0.000001))).^m;
end
%采用二进制Gray编码,其目的是为了克服二进制编码的Hamming悬崖缺点
function[initpop]=InitPopGray(popsize,bits)
len=sum(bits);
initpop=zeros(popsize,len);
%Thewholezeroencodingindividual
fori=2:
popsize-1
pop=round(rand(1,len));
pop=mod(([0pop]+[pop0]),2);
%i=1时,b
(1)=a
(1);
i>
1时,b(i)=mod(a(i-1)+a(i),2)
%其中原二进制串:
a
(1)a
(2)...a(n),Gray串:
b
(1)b
(2)...b(n)
initpop(i,:
)=pop(1:
end-1);
initpop(popsize,:
)=ones(1,len);
%Thewholeoneencodingindividual
%解码
function[fval]=b2f(bval,bounds,bits)
%fval-表征各变量的十进制数
%bval-表征各变量的二进制编码串
%bounds-各变量的取值范围
%bits-各变量的二进制编码长度
scale=(bounds(:
./(2.^bits-1);
%Therangeofthevariables
numV=size(bounds,1);
cs=[0cumsum(bits)];
numV
a=bval((cs(i)+1):
cs(i+1));
fval(i)=sum(2.^(size(a,2)-1:
-1:
0).*a)*scale(i)+bounds(i,1);
%选择操作
%采用基于轮盘赌法的非线性排名选择
%各个体成员按适应值从大到小分配选择概率:
%P(i)=(q/1-(1-q)^n)*(1-q)^i,其中P(0)>
P
(1)>
...>
P(n),sum(P(i))=1
function[selectpop]=NonlinearRankSelect(FUN,pop,bounds,bits)
globalmn
selectpop=zeros(m,n);
fit=zeros(m,1);
fit(i)=feval(FUN(1,:
),(b2f(pop(i,:
%以函数值为适应值做排名依据
selectprob=fit/sum(fit);
%计算各个体相对适应度(0,1)
q=max(selectprob);
%选择最优的概率
x=zeros(m,2);
x(:
1)=[m:
1]'
[yx(:
2)]=sort(selectprob);
r=q/(1-(1-q)^m);
%标准分布基值
newfit(x(:
2))=r*(1-q).^(x(:
1)-1);
%生成选择概率
newfit=cumsum(newfit);
%计算各选择概率之和
rNums=sort(rand(m,1));
fitIn=1;
new