基于python和MATLAB的遗传算法优化函数最小值

遗传算法计算函数最小值f（x）=∑xi^2 基于matlib与python对比

采用种群数100，实数编码，进化100代交叉概率0.8，变异概率0.1

# -*- coding: utf-8 -*-
"""

"""


import numpy as np



#适应度函数
def func2(x):
    summ=sum(pow(x,2))
    return summ

D=10
NP=100
Xs=20
Xx=-20
G=100
Pc=0.8
Pm=0.1
nf=np.zeros((D,NP))
f=np.random.rand(D,NP)*(Xs-Xx)+Xx
Sortf=np.zeros((D,NP))
NSortf=np.zeros((D,NP))
MSLL=np.zeros(NP)
NMSLL=np.zeros(NP)
NMSLL1=np.zeros(NP)
#Index=np.zeros(NP)
trace=[]
for w in range(NP):          
    MSLL[w]=func2(f[:,w])     #MSLL是f的适应度
Index=np.argsort(MSLL)
SortMSLL=sorted(MSLL)           #SortMSLL是从低到高排序后的MSLL


Sortf=f[:,Index]             #由f升序得到

for gen in range(G):
    Emper=Sortf[:,0]          #选择君主
    NoPoint=round(D*Pc)      #交叉点个数
    PoPoint=np.random.randint(0,D,[NoPoint,int(NP/2)])
    nf=Sortf
    for i in range(int(NP/2)):
        nf[:,2*i]=Emper
        nf[:,2*i+1]=Sortf[:,2*i+1]
        for k in range(NoPoint):
            nf[PoPoint[k,i],2*i]=nf[PoPoint[k,i],2*i+1]
            nf[PoPoint[k,i],2*i+1]=Emper[PoPoint[k,i]]
    #变异操作
    for by in range(NP):
        for i in range(D):
            r=np.random.rand()
            if r<Pm:
                nf[i,by]=np.random.rand()*(Xs-Xx)+Xx
    #子群按适应度排序
                
    for i in range(NP):
        NMSLL[i]=func2(nf[:,i])
    Index=np.argsort(NMSLL)
    NSortMSLL=sorted(NMSLL)
    
    NSortf=nf[:,Index]
    f1=np.hstack((Sortf,NSortf))
    MSLL1=np.hstack((SortMSLL,NSortMSLL))
    Index=np.argsort(MSLL1)
    SortMSLL1=sorted(MSLL1)
    
    Sortf1=f1[:,Index]
    SortMSLL=SortMSLL1[0:NP]
    Sortf=Sortf1[:,0:NP]
    trace.append(SortMSLL[0])
    
    
    
	MATLIB

```python
clear all
close all
clc




D=10
NP=100
Xs=20
Xx=-20
G=25
f=zeros(D,NP)
nf=zeros(D,NP)
Pc=0.8
Pm=0.1
f=rand(D,NP)*(Xs-Xx)+Xx
for np=1:NP
    MSLL(np)=func2(f(:,np))
end
[SortMSLL,Index]=sort(MSLL)
Sortf=f(:,Index)

%%%%君主方式进行选择交叉
for gen=1:G
    Emper=Sortf(:,1)   %君主染色体
    NoPoint=round(D*Pc)
    PoPoint=randi([1,D],NoPoint,NP/2)
    nf=Sortf
    for i=1:NP/2
        nf(:,2*i-1)=Emper
        nf(:,2*i)=Sortf(:,2*i)
        for k=1:NoPoint
            nf(PoPoint(k,i),2*i-1)=nf(PoPoint(k,i),2*i)
            nf(PoPoint(k,i),2*i)=Emper(PoPoint(k,i))
        end
    end
    for m =1:NP
        for n=1:D
            r=rand(1,1)
            if r<Pm
                nf(n,m)=rand(1,1)*(Xs-Xx)+Xx
            end
        end
    end
    for np=1:NP
        NMSLL(np)=func2(nf(:,np))
    end
    [NSortMSLL,Index]=sort(NMSLL)
    NSortf=nf(:,Index)
    f1=[Sortf,NSortf]
    MSLL1=[SortMSLL,NSortMSLL]
    [SortMSLL1,Index]=sort(MSLL1)
    Sortf1=f1(:,Index)
    SortMSLL=SortMSLL1(1:NP)
    Sortf=Sortf1(:,1:NP)
    trace(gen)=SortMSLL(1)
end
bestf=Sortf(:,1)
trace(end)
figure
plot(trace)
xlabel('迭代次数')
ylabel('目标函数值')



function result=func2(x)
sumn=sum(x.^2)
result=sumn
end