系列博客：
为了解决参数优化问题，学习以下三种智能优化算法：
1.免疫算法
2.蚁群优化算法
3.粒子群算法
一、三种算法实例学习
1.1 免疫算法
1.1.1 代码部分
%%%%%%%%%%%%%%%%%免疫算法求函数极值%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%初始化%%%%%%%%%%%%%%%%%%%%%%%%%
clear all; %清除所有变量
close all; %清图
clc; %清屏
D=10; %免疫个体维数
NP=100; %免疫个体数目
Xs=20; %取值上限
Xx=-20; %取值下限
G=500; %最大免疫代数
pm=0.7; %变异概率
alfa=1; %激励度系数
belta=1; %激励度系数
detas=0.2; %相似度阈值
gen=0; %免疫代数
Ncl=10; %克隆个数
deta0=1*Xs; %邻域范围初值
%%%%%%%%%%%%%%%%%%%%%%%初始种群%%%%%%%%%%%%%%%%%%%%%%%%
f=rand(D,NP)*(Xs-Xx)+Xx;
for np=1:NP
MSLL(np)=mianyi_func1(f(:,np));
end
%%%%%%%%%%%%%%%%%计算个体浓度和激励度%%%%%%%%%%%%%%%%%%%
for np=1:NP
for j=1:NP
nd(j)=sum(sqrt((f(:,np)-f(:,j)).^2));
if nd(j)<detas
nd(j)=1;
else
nd(j)=0;
end
end
ND(np)=sum(nd)/NP;
end
MSLL = alfa*MSLL- belta*ND;
%%%%%%%%%%%%%%%%%%%激励度按升序排列%%%%%%%%%%%%%%%%%%%%%%
[SortMSLL,Index]=sort(MSLL);
Sortf=f(:,Index);
%%%%%%%%%%%%%%%%%%%%%%%%免疫循环%%%%%%%%%%%%%%%%%%%%%%%%
while gen<G
for i=1:NP/2
%%%%%%%%选激励度前NP/2个体进行免疫操作%%%%%%%%%%%
a=Sortf(:,i);
Na=repmat(a,1,Ncl);
deta=deta0/(gen+0.0001);
for j=1:Ncl
for ii=1:D
%%%%%%%%%%%%%%%%%变异%%%%%%%%%%%%%%%%%%%
if rand<pm
Na(ii,j)=Na(ii,j)+(rand-0.5)*deta;
end
%%%%%%%%%%%%%%边界条件处理%%%%%%%%%%%%%%%
if (Na(ii,j)>Xs) | (Na(ii,j)<Xx)
Na(ii,j)=rand * (Xs-Xx)+Xx;
end
end
end
Na(:,1)=Sortf(:,i); %保留克隆源个体
%%%%%%%%%%克隆抑制，保留亲和度最高的个体%%%%%%%%%%
for j=1:Ncl
NaMSLL(j)=mianyi_func1(Na(:,j));
end
[NaSortMSLL,Index]=sort(NaMSLL);
aMSLL(i)=NaSortMSLL(1);
NaSortf=Na(:,Index);
af(:,i)=NaSortf(:,1);
end
%%%%%%%%%%%%%%%%%%%%免疫种群激励度%%%%%%%%%%%%%%%%%%%
for np=1:NP/2
for j=1:NP/2
nda(j)=sum(sqrt((af(:,np)-af(:,j)).^2));
if nda(j)<detas
nda(j)=1;
else
nda(j)=0;
end
end
aND(np)=sum(nda)/NP/2;
end
aMSLL = alfa*aMSLL- belta*aND;
%%%%%%%%%%%%%%%%%%%%%%%种群刷新%%%%%%%%%%%%%%%%%%%%%%%
bf=rand(D,NP/2)*(Xs-Xx)+Xx;
for np=1:NP/2
bMSLL(np)=mianyi_func1(bf(:,np));
end
%%%%%%%%%%%%%%%%%%%新生成种群激励度%%%%%%%%%%%%%%%%%%%%
for np=1:NP/2
for j=1:NP/2
ndc(j)=sum(sqrt((bf(:,np)-bf(:,j)).^2));
if ndc(j)<detas
ndc(j)=1;
else
ndc(j)=0;
end
end
bND(np)=sum(ndc)/NP/2;
end
bMSLL = alfa*bMSLL- belta*bND;
%%%%%%%%%%%%%%免疫种群与新生种群合并%%%%%%%%%%%%%%%%%%%
f1=[af,bf];
MSLL1=[aMSLL,bMSLL];
[SortMSLL,Index]=sort(MSLL1);
Sortf=f1(:,Index);
gen=gen+1;
trace(gen)=mianyi_func1(Sortf(:,1));
end
%%%%%%%%%%%%%%%%%%%%%%%输出优化结果%%%%%%%%%%%%%%%%%%%%%%%%
Bestf=Sortf(:,1); %最优变量
trace(end); %最优值
figure,plot(trace)
xlabel('迭代次数')
ylabel('目标函数值')
title('亲和度进化曲线')
%%%%%%%%%%%%%%%%%%%%%%%%%亲和度函数%%%%%%%%%%%%%%%%%%%%%%
function result=mianyi_func1(x)
summ=sum(x.^2);
result=summ;
1.1.2 运行结果
1.1.3 相关分析
关键参数分析：
（1）抗体种群大小NP
抗体种群保留了免疫细胞的多样性，从直观上看，种群越大，免疫算法的全局搜索能力越好，但是算法每代的计算量也相应增大。在大多数问题中，NP取10~100较为合适，一般不超过200。
（2）免疫选择比例
克隆选择的抗体的数量越多，将产生更多的克隆，其搜索能力越强，但是将增加每代的计算量。一般可以取抗体种群大小NP的10%~50%。
（3）抗体克隆扩增的倍数
克隆的倍数决定了克隆扩增的细胞的数量，从而决定了算法的搜索能力，主要是局部搜索能力。克隆倍数数值越大，局部搜索能力越好，全局搜索能力也有一定提高，但是计算量也随之增大，一般取5~10倍。
（4）种群刷新比例
z细胞的淘汰和更新时产生抗体多样性的重要机制，因而对免疫算法的全局搜索能力产生重要影响。每代更新的抗体一般不超过抗体种群的50%。
（5）最大进化代数G
最大进化代数G时表示免疫算法运行结束条件的一个参数，表示免疫算法运行到指定的进化代数之后就停止运行，并将当前全体中的最佳个体作为所求问题的最优解输出。一般G取100~500。
1.2 蚁群算法
1.2.1 代码部分
%%%%%%%%%%%%%%%%%%%%蚁群算法解决TSP问题%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%初始化%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
clear all; %清除所有变量
close all; %清图
clc; %清屏
m=50; %蚂蚁个数
Alpha=1; %信息素重要程度参数
Beta=5; %启发式因子重要程度参数
Rho=0.1; %信息素蒸发系数
G_max=200; %最大迭代次数
Q=100; %信息素增加强度系数
C=[6734 1453
2233 10
5530 1424
401 841
3082 1644
7608 4458
7573 3716
7265 1268
6898 1885
1112 2049
5468 2606
5989 2873
4706 2674
4612 2035
6347 2683
6107 669
7611 5184
7462 3590
7732 4723
5900 3561
4483 3369
6101 1110
5199 2182
1633 2809
4307 2322
675 1006
7555 4819
7541 3981
3177 756
7352 4506
7545 2801
3245 3305
6426 3173
4608 1198
23 2216
7248 3779
7762 4595
7392 2244
3484 2829
6271 2135
4985 140
1916 1569
7280 4899
7509 3239
10 2676
6807 2993
5185 3258
3023 1942]; %31个省会城市坐标
%%%%%%%%%%%%%%%%%%%%%%%%第一步：变量初始化%%%%%%%%%%%%%%%%%%%%%%%%
n=size(C,1); %n表示问题的规模（城市个数）
D=zeros(n,n); %D表示两个城市距离间隔矩阵
for i=1:n
for j=1:n
if i~=j
D(i,j)=((C(i,1)-C(j,1))^2+(C(i,2)-C(j,2))^2)^0.5;
else
D(i,j)=eps;
end
D(j,i)=D(i,j);
end
end
Eta=1./D; %Eta为启发因子，这里设为距离的倒数
Tau=ones(n,n); %Tau为信息素矩阵
Tabu=zeros(m,n); %存储并记录路径的生成
NC=1; %迭代计数器
R_best=zeros(G_max,n); %各代最佳路线
L_best=inf.*ones(G_max,1); %各代最佳路线的长度
figure(1);%优化解
while NC<=G_max
%%%%%%%%%%%%%%%%%%第二步：将m只蚂蚁放到n个城市上%%%%%%%%%%%%%%%%
Randpos=[];
for i=1:(ceil(m/n))
Randpos=[Randpos,randperm(n)];
end
Tabu(:,1)=(Randpos(1,1:m))';
%%%%%第三步：m只蚂蚁按概率函数选择下一座城市，完成各自的周游%%%%%%
for j=2:n
for i=1:m
visited=Tabu(i,1:(j-1)); %已访问的城市
J=zeros(1,(n-j+1)); %待访问的城市
P=J; %待访问城市的选择概率分布
Jc=1;
for k=1:n
if length(find(visited==k))==0
J(Jc)=k;
Jc=Jc+1;
end
end
%%%%%%%%%%%%%%%%%%计算待选城市的概率分布%%%%%%%%%%%%%%%%
for k=1:length(J)
P(k)=(Tau(visited(end),J(k))^Alpha)...
*(Eta(visited(end),J(k))^Beta);
end
P=P/(sum(P));
%%%%%%%%%%%%%%%%按概率原则选取下一个城市%%%%%%%%%%%%%%%%
Pcum=cumsum(P);
Select=find(Pcum>=rand);
to_visit=J(Select(1));
Tabu(i,j)=to_visit;
end
end
if NC>=2
Tabu(1,:)=R_best(NC-1,:);
end
%%%%%%%%%%%%%%%%%%%第四步：记录本次迭代最佳路线%%%%%%%%%%%%%%%%%%
L=zeros(m,1);
for i=1:m
R=Tabu(i,:);
for j=1:(n-1)
L(i)=L(i)+D(R(j),R(j+1));
end
L(i)=L(i)+D(R(1),R(n));
end
L_best(NC)=min(L);
pos=find(L==L_best(NC));
R_best(NC,:)=Tabu(pos(1),:);
%%%%%%%%%%%%%%%%%%%%%%%%%第五步：更新信息素%%%%%%%%%%%%%%%%%%%%%%
Delta_Tau=zeros(n,n);
for i=1:m
for j=1:(n-1)
Delta_Tau(Tabu(i,j),Tabu(i,j+1))=...
Delta_Tau(Tabu(i,j),Tabu(i,j+1))+Q/L(i);
end
Delta_Tau(Tabu(i,n),Tabu(i,1))=...
Delta_Tau(Tabu(i,n),Tabu(i,1))+Q/L(i);
end
Tau=(1-Rho).*Tau+Delta_Tau;
%%%%%%%%%%%%%%%%%%%%%%%第六步：禁忌表清零%%%%%%%%%%%%%%%%%%%%%%
Tabu=zeros(m,n);
%%%%%%%%%%%%%%%%%%%%%%%%%历代最优路线%%%%%%%%%%%%%%%%%%%%%%%%%%
for i=1:n-1
plot([ C(R_best(NC,i),1), C(R_best(NC,i+1),1)],...
[C(R_best(NC,i),2), C(R_best(NC,i+1),2)],'bo-');
hold on;
end
plot([C(R_best(NC,n),1), C(R_best(NC,1),1)],...
[C(R_best(NC,n),2), C(R_best(NC,1),2)],'ro-');
title(['优化最短距离:',num2str(L_best(NC))]);
hold off;
pause(0.005);
NC=NC+1;
end
%%%%%%%%%%%%%%%%%%%%%%%%%%第七步：输出结果%%%%%%%%%%%%%%%%%%%%%%%%%%
Pos=find(L_best==min(L_best));
Shortest_Route=R_best(Pos(1),:); %最佳路线
Shortest_Length=L_best(Pos(1)); %最佳路线长度
figure(2),
plot(L_best)
xlabel('迭代次数')
ylabel('目标函数值')
title('适应度进化曲线')1.2.2 运行结果
1.3 粒子群问题
1.3.1 代码部分
%%%%%%%%%%%%%%%%%粒子群算法求函数极值%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%初始化%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
clear all; %清除所有变量
close all; %清图
clc; %清屏
N=100; %群体粒子个数
D=10; %粒子维数
T=200; %最大迭代次数
c1=1.5; %学习因子1
c2=1.5; %学习因子2
w=0.8; %惯性权重
Xmax=20; %位置最大值
Xmin=-20; %位置最小值
Vmax=10; %速度最大值
Vmin=-10; %速度最小值
%%%%%%%%%%%%%%%%初始化种群个体（限定位置和速度）%%%%%%%%%%%%%%%%
x=rand(N,D) * (Xmax-Xmin)+Xmin;
v=rand(N,D) * (Vmax-Vmin)+Vmin;
%%%%%%%%%%%%%%%%%%初始化个体最优位置和最优值%%%%%%%%%%%%%%%%%%%
p=x;
pbest=ones(N,1);
for i=1:N
pbest(i)=liziqun_func1(x(i,:));
end
%%%%%%%%%%%%%%%%%%%初始化全局最优位置和最优值%%%%%%%%%%%%%%%%%%
g=ones(1,D);
gbest=inf;
for i=1:N
if(pbest(i)<gbest)
g=p(i,:);
gbest=pbest(i);
end
end
gb=ones(1,T);
%%%%%%%%%%%按照公式依次迭代直到满足精度或者迭代次数%%%%%%%%%%%%%
for i=1:T
for j=1:N
%%%%%%%%%%%%%%更新个体最优位置和最优值%%%%%%%%%%%%%%%%%
if (liziqun_func1(x(j,:))<pbest(j))
p(j,:)=x(j,:);
pbest(j)=liziqun_func1(x(j,:));
end
%%%%%%%%%%%%%%%%更新全局最优位置和最优值%%%%%%%%%%%%%%%
if(pbest(j)<gbest)
g=p(j,:);
gbest=pbest(j);
end
%%%%%%%%%%%%%%%%%跟新位置和速度值%%%%%%%%%%%%%%%%%%%%%
v(j,:)=w*v(j,:)+c1*rand*(p(j,:)-x(j,:))...
+c2*rand*(g-x(j,:));
x(j,:)=x(j,:)+v(j,:);
%%%%%%%%%%%%%%%%%%%%边界条件处理%%%%%%%%%%%%%%%%%%%%%%
for ii=1:D
if (v(j,ii)>Vmax) | (v(j,ii)< Vmin)
v(j,ii)=rand * (Vmax-Vmin)+Vmin;
end
if (x(j,ii)>Xmax) | (x(j,ii)< Xmin)
x(j,ii)=rand * (Xmax-Xmin)+Xmin;
end
end
end
%%%%%%%%%%%%%%%%%%%%记录历代全局最优值%%%%%%%%%%%%%%%%%%%%%
gb(i)=gbest;
end
g; %最优个体
gb(end); %最优值
figure
plot(gb)
xlabel('迭代次数');
ylabel('适应度值');
title('适应度进化曲线')
%%%%%%%%%%%%%%%%%%%适应度函数%%%%%%%%%%%%%%%%%%%%
function result=liziqun_func1(x)
summ=sum(x.^2);
result=summ;
1.3.2 运行结果
二、三种算法比较
以求解函数f(x,y)=3cos(xy)+x+y^2的最小值进行比较，其中x的取值范围为[-4,4],y的取值范围为[-4,-4]。
f(x,y)是一个有多个局部极值的函数，其函数图像如图所示：
画图代码如下：
clear all;
close all;
clc;
x=-4:0.02:4;
y=-4:0.02:4;
N=size(x,2);
for i=1:N
for j=1:N
z(i,j)=3*cos(x(i)*y(j))+x(i)+y(j).^2;
end
end
mesh(x,y,z)
xlabel('x')
ylabel('y')2.1 免疫算法求解函数极值
2.1.1 代码部分
%%%%%%%%%%%%%%%%%免疫算法求函数极值%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%初始化%%%%%%%%%%%%%%%%%%%%%%%%%
clear all; %清除所有变量
close all; %清图
clc; %清屏
D=2; %免疫个体维数
NP=100; %免疫个体数目
Xs=4; %取值上限
Xx=-4; %取值下限
G=500; %最大免疫代数
pm=0.7; %变异概率
alfa=1; %激励度系数
belta=1; %激励度系数
detas=0.2; %相似度阈值
gen=0; %免疫代数
Ncl=10; %克隆个数
deta0=1*Xs; %邻域范围初值
%%%%%%%%%%%%%%%%%%%%%%%初始种群%%%%%%%%%%%%%%%%%%%%%%%%
f=rand(D,NP)*(Xs-Xx)+Xx;
for np=1:NP
MSLL(np)=func(f(:,np));
end
%%%%%%%%%%%%%%%%%计算个体浓度和激励度%%%%%%%%%%%%%%%%%%%
for np=1:NP
for j=1:NP
nd(j)=sum(sqrt((f(:,np)-f(:,j)).^2));
if nd(j)<detas
nd(j)=1;
else
nd(j)=0;
end
end
ND(np)=sum(nd)/NP;
end
MSLL = alfa*MSLL- belta*ND;
%%%%%%%%%%%%%%%%%%%激励度按升序排列%%%%%%%%%%%%%%%%%%%%%%
[SortMSLL,Index]=sort(MSLL);
Sortf=f(:,Index);
%%%%%%%%%%%%%%%%%%%%%%%%免疫循环%%%%%%%%%%%%%%%%%%%%%%%%
while gen<G
for i=1:NP/2
%%%%%%%%选激励度前NP/2个体进行免疫操作%%%%%%%%%%%
a=Sortf(:,i);
Na=repmat(a,1,Ncl);
deta=deta0/(gen+0.0001);
for j=1:Ncl
for ii=1:D
%%%%%%%%%%%%%%%%%变异%%%%%%%%%%%%%%%%%%%
if rand<pm
Na(ii,j)=Na(ii,j)+(rand-0.5)*deta;
end
%%%%%%%%%%%%%%边界条件处理%%%%%%%%%%%%%%%
if (Na(ii,j)>Xs) | (Na(ii,j)<Xx)
Na(ii,j)=rand * (Xs-Xx)+Xx;
end
end
end
Na(:,1)=Sortf(:,i); %保留克隆源个体
%%%%%%%%%%克隆抑制，保留亲和度最高的个体%%%%%%%%%%
for j=1:Ncl
NaMSLL(j)=func(Na(:,j));
end
[NaSortMSLL,Index]=sort(NaMSLL);
aMSLL(i)=NaSortMSLL(1);
NaSortf=Na(:,Index);
af(:,i)=NaSortf(:,1);
end
%%%%%%%%%%%%%%%%%%%%免疫种群激励度%%%%%%%%%%%%%%%%%%%
for np=1:NP/2
for j=1:NP/2
nda(j)=sum(sqrt((af(:,np)-af(:,j)).^2));
if nda(j)<detas
nda(j)=1;
else
nda(j)=0;
end
end
aND(np)=sum(nda)/NP/2;
end
aMSLL = alfa*aMSLL- belta*aND;
%%%%%%%%%%%%%%%%%%%%%%%种群刷新%%%%%%%%%%%%%%%%%%%%%%%
bf=rand(D,NP/2)*(Xs-Xx)+Xx;
for np=1:NP/2
bMSLL(np)=func(bf(:,np));
end
%%%%%%%%%%%%%%%%%%%新生成种群激励度%%%%%%%%%%%%%%%%%%%%
for np=1:NP/2
for j=1:NP/2
ndc(j)=sum(sqrt((bf(:,np)-bf(:,j)).^2));
if ndc(j)<detas
ndc(j)=1;
else
ndc(j)=0;
end
end
bND(np)=sum(ndc)/NP/2;
end
bMSLL = alfa*bMSLL- belta*bND;
%%%%%%%%%%%%%%免疫种群与新生种群合并%%%%%%%%%%%%%%%%%%%
f1=[af,bf];
MSLL1=[aMSLL,bMSLL];
[SortMSLL,Index]=sort(MSLL1);
Sortf=f1(:,Index);
gen=gen+1;
trace(gen)=func(Sortf(:,1));
end
%%%%%%%%%%%%%%%%%%%%%%%输出优化结果%%%%%%%%%%%%%%%%%%%%%%%%
Bestf=Sortf(:,1); %最优变量
trace(end); %最优值
figure,plot(trace)
xlabel('迭代次数')
ylabel('目标函数值')
title('亲和度进化曲线')
%%%%%%%%%%%%%%%%%%%%%%%%%目标函数%%%%%%%%%%%%%%%%%%%%%%
function result=func(x)
result=3*cos(x(1)*x(2))+x(1)+x(2).^2;
end
2.1.2 结果分析
在免疫代数为500，即x=-4,y=0.7539时，函数取得最小值-6.40785.
免疫算法不强调算法参数设置和初始解的质量，利用其启发式的智能搜索机制，即使起步于劣质解种群，最终也可以搜索到问题的全局最优解，对问题和初始解的依赖性不强，具有很强的适应性和鲁棒性。故在上图，种群20代以后基本处于最优解。
2.2 蚁群算法求解函数极值
2.2.1 代码部分
%%%%%%%%%%%%%%%%%%%%蚁群算法求函数极值%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%初始化%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
clear all; %清除所有变量
close all; %清图
clc; %清屏
m=20; %蚂蚁个数
G_max=200; %最大迭代次数
Rho=0.9; %信息素蒸发系数
P0=0.2; %转移概率常数
XMAX= 4; %搜索变量x最大值
XMIN= -4; %搜索变量x最小值
YMAX= 4; %搜索变量y最大值
YMIN= -4; %搜索变量y最小值
%%%%%%%%%%%%%%%%%随机设置蚂蚁初始位置%%%%%%%%%%%%%%%%%%%%%%
for i=1:m
X(i,1)=(XMIN+(XMAX-XMIN)*rand);
X(i,2)=(YMIN+(YMAX-YMIN)*rand);
Tau(i)=func1(X(i,1),X(i,2));
end
step=0.1; %局部搜索步长
for NC=1:G_max
lamda=1/NC;
[Tau_best,BestIndex]=min(Tau);
%%%%%%%%%%%%%%%%%%计算状态转移概率%%%%%%%%%%%%%%%%%%%%
for i=1:m
P(NC,i)=(Tau(BestIndex)-Tau(i))/Tau(BestIndex);
end
%%%%%%%%%%%%%%%%%%%%%%位置更新%%%%%%%%%%%%%%%%%%%%%%%%
for i=1:m
%%%%%%%%%%%%%%%%%局部搜索%%%%%%%%%%%%%%%%%%%%%%
if P(NC,i)<P0
temp1=X(i,1)+(2*rand-1)*step*lamda;
temp2=X(i,2)+(2*rand-1)*step*lamda;
else
%%%%%%%%%%%%%%%%全局搜索%%%%%%%%%%%%%%%%%%%%%%%
temp1=X(i,1)+(XMAX-XMIN)*(rand-0.5);
temp2=X(i,2)+(YMAX-YMIN)*(rand-0.5);
end
%%%%%%%%%%%%%%%%%%%%%边界处理%%%%%%%%%%%%%%%%%%%%%%%
if temp1<XMIN
temp1=XMIN;
end
if temp1>XMAX
temp1=XMAX;
end
if temp2<YMIN
temp2=YMIN;
end
if temp2>YMAX
temp2=YMAX;
end
%%%%%%%%%%%%%%%%%%蚂蚁判断是否移动%%%%%%%%%%%%%%%%%%
if func1(temp1,temp2)<func1(X(i,1),X(i,2))
X(i,1)=temp1;
X(i,2)=temp2;
end
end
%%%%%%%%%%%%%%%%%%%%%%%更新信息素%%%%%%%%%%%%%%%%%%%%%%%
for i=1:m
Tau(i)=(1-Rho)*Tau(i)+func1(X(i,1),X(i,2));
end
[value,index]=min(Tau);
trace(NC)=func1(X(index,1),X(index,2));
end
[min_value,min_index]=min(Tau);
minX=X(min_index,1) %最优变量
minY=X(min_index,2) %最优变量
minValue=func1(X(min_index,1),X(min_index,2)); %最优值
figure
plot(trace)
xlabel('搜索次数');
ylabel('适应度值');
title('适应度进化曲线')
%%%%%%%%%%%目标函数%%%%%%%%%%%
function value=func1(x,y)
value =3*cos(x*y)+x+y.^2;
end2.2.2 结果分析
在200轮迭代，即x=-4,y=-0.7539时，函数取得最小值-6.40786.
蚁群算法一般需要较长的搜索时间和容易出现停滞现象等不足，故在上图中，容易看到，在80代后，才基本找到最优解。
2.3 粒子群算法求解函数极值
2.3.1 代码部分
%%%%%%%%%%%%%%%%%粒子群算法求函数极值%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%初始化%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
clear all; %清除所有变量
close all; %清图
clc; %清屏
N=100; %群体粒子个数
D=10; %粒子维数
T=200; %最大迭代次数
c1=1.5; %学习因子1
c2=1.5; %学习因子2
w=0.8; %惯性权重
Xmax=4; %位置最大值
Xmin=-4; %位置最小值
Vmax=1; %速度最大值
Vmin=-1; %速度最小值
%%%%%%%%%%%%%%%%初始化种群个体（限定位置和速度）%%%%%%%%%%%%%%%%
x=rand(N,D) * (Xmax-Xmin)+Xmin;
v=rand(N,D) * (Vmax-Vmin)+Vmin;
%%%%%%%%%%%%%%%%%%初始化个体最优位置和最优值%%%%%%%%%%%%%%%%%%%
p=x;
pbest=ones(N,1);
for i=1:N
pbest(i)=func(x(i,:));
end
%%%%%%%%%%%%%%%%%%%初始化全局最优位置和最优值%%%%%%%%%%%%%%%%%%
g=ones(1,D);
gbest=inf;
for i=1:N
if(pbest(i)<gbest)
g=p(i,:);
gbest=pbest(i);
end
end
gb=ones(1,T);
%%%%%%%%%%%按照公式依次迭代直到满足精度或者迭代次数%%%%%%%%%%%%%
for i=1:T
for j=1:N
%%%%%%%%%%%%%%更新个体最优位置和最优值%%%%%%%%%%%%%%%%%
if (func(x(j,:))<pbest(j))
p(j,:)=x(j,:);
pbest(j)=func(x(j,:));
end
%%%%%%%%%%%%%%%%更新全局最优位置和最优值%%%%%%%%%%%%%%%
if(pbest(j)<gbest)
g=p(j,:);
gbest=pbest(j);
end
%%%%%%%%%%%%%%%%%跟新位置和速度值%%%%%%%%%%%%%%%%%%%%%
v(j,:)=w*v(j,:)+c1*rand*(p(j,:)-x(j,:))...
+c2*rand*(g-x(j,:));
x(j,:)=x(j,:)+v(j,:);
%%%%%%%%%%%%%%%%%%%%边界条件处理%%%%%%%%%%%%%%%%%%%%%%
for ii=1:D
if (v(j,ii)>Vmax) | (v(j,ii)< Vmin)
v(j,ii)=rand * (Vmax-Vmin)+Vmin;
end
if (x(j,ii)>Xmax) | (x(j,ii)< Xmin)
x(j,ii)=rand * (Xmax-Xmin)+Xmin;
end
end
end
%%%%%%%%%%%%%%%%%%%%记录历代全局最优值%%%%%%%%%%%%%%%%%%%%%
gb(i)=gbest;
end
g %最优个体
gb(end) %最优值
figure
plot(gb)
xlabel('迭代次数');
ylabel('适应度值');
title('适应度进化曲线')
%%%%%%%%%%%%%%%%%%%%%%%%%目标函数%%%%%%%%%%%%%%%%%%%%%%
function result=func(x)
result=3*cos(x(1)*x(2))+x(1)+x(2).^2;
end
2.3.2 结果分析
在200轮迭代，函数取得最小值-6.40704.
粒子群算法本质是一种随机搜索算法，它是一种新型的智能优化技术。该算法能以较大概率收敛于全局最优解。如上图所示，算法在70次迭代，基本找到全局最优解。实践证明，粒子群算法适合在动态、多目标优化环境中寻优，与传统优化算法相比，具有较快的计算速度和更好地全局搜索能力。所以，与其他算法相比，粒子群算法是一种高效的并行搜索算法。
605
被折叠的 条评论
为什么被折叠?
到【灌水乐园】发言
