学习代码来源于：遗传算法python
# -*-coding:utf-8 -*-
#目标求解sin(x)最大值
import random
import math
import matplotlib.pyplot as plt
#初始化种群 生成chromosome_length大小的population_size个个体的种群
def species_origin(population_size,chromosome_length):
population=[[]]
#one dimension represent a individual
for i in range(population_size):
temporary=[]
#染色体暂存器
for j in range(chromosome_length):
temporary.append(random.randint(0,1))
#随机产生一个染色体,由二进制数组成
population.append(temporary)
#将染色体添加到种群中
return population[1:]
# 将种群返回，种群是个二维数组，个体和染色体两维
#从二进制到十进制
#编码 input:种群,染色体长度
def translation(population,chromosome_length):
temporary=[]
for i in range(len(population)):
total=0
for j in range(chromosome_length):
total+=population[i][j]*(math.pow(2,j))
#从第一个基因开始，每位对2求幂，再求和
# 如：0101 转成十进制为：1 * 20 + 0 * 21 + 1 * 22 + 0 * 23 = 1 + 0 + 4 + 0 = 5
temporary.append(total)
#一个染色体编码完成，由一个二进制数编码为一个十进制数
return temporary
# 返回种群中所有个体编码完成后的十进制数
# 目标函数相当于环境 对染色体进行筛选，这里是sin(x)
def function(population,chromosome_length,max_value):
temporary=[]
function1=[]
temporary=translation(population,chromosome_length)
# 暂存种群中的所有的染色体(十进制)
for i in range(len(temporary)):
x=temporary[i]*max_value/(math.pow(2,chromosome_length)-1)
#一个基因代表一个决策变量，其算法是先转化成十进制，然后再除以2的基因个数次方减1(固定值)。
function1.append(math.sin(x))
#这里将sin(x)作为目标函数
return function1
#筛去不可能的适应度
def fitness(function1):
fitness1=[]
min_fitness=mf=0
for i in range(len(function1)):
if(function1[i]+mf>0):
temporary=mf+function1[i]
else:
temporary=0.0
# 如果适应度小于0,则定为0
fitness1.append(temporary)
#将适应度添加到列表中
return fitness1
#计算适应度和
def sum(fitness1):
total=0
for i in range(len(fitness1)):
total+=fitness1[i]
return total
#计算适应度斐伯纳且列表，目的是求出累积概率。
def cumsum(fitness1):
for i in range(len(fitness1)-2,-1,-1):
# range(start,stop,[step])
# 倒计数
total=0
j=0
while(j<=i):
total+=fitness1[j]
j+=1
fitness1[i]=total
fitness1[len(fitness1)-1]=1
#3.选择种群中个体适应度最大的个体
def selection(population,fitness1):
new_fitness=[]
#单个公式暂存器
total_fitness=sum(fitness1)
#将所有的适应度求和
for i in range(len(fitness1)):
new_fitness.append(fitness1[i]/total_fitness)
#将所有个体的适应度正则化
cumsum(new_fitness)
#
ms=[]
#存活的种群
population_length=pop_len=len(population)
#求出种群长度
#根据随机数确定哪几个能存活
for i in range(pop_len):
ms.append(random.random())
# 产生种群个数的随机值
ms.sort()
# 存活的种群排序
fitin=0
newin=0
new_population=new_pop=population
#轮盘赌方式
while newin<pop_len:
if(ms[newin]<new_fitness[fitin]):
new_pop[newin]=pop[fitin]
newin+=1
else:
fitin+=1
population=new_pop
#4.交叉操作
def crossover(population,pc):
#pc是概率阈值，选择单点交叉还是多点交叉，生成新的交叉个体，这里没用
pop_len=len(population)
for i in range(pop_len-1):
cpoint=random.randint(0,len(population[0]))
#在种群个数内随机生成单点交叉点
temporary1=[]
temporary2=[]
temporary1.extend(pop[i][0:cpoint])
temporary1.extend(pop[i+1][cpoint:len(population[i])])
#将tmporary1作为暂存器，暂时存放第i个染色体中的前0到cpoint个基因，
#然后再把第i+1个染色体中的后cpoint到第i个染色体中的基因个数，补充到temporary2后面
temporary2.extend(pop[i+1][0:cpoint])
temporary2.extend(pop[i][cpoint:len(pop[i])])
# 将tmporary2作为暂存器，暂时存放第i+1个染色体中的前0到cpoint个基因，
# 然后再把第i个染色体中的后cpoint到第i个染色体中的基因个数，补充到temporary2后面
pop[i]=temporary1
pop[i+1]=temporary2
# 第i个染色体和第i+1个染色体基因重组/交叉完成
#step4：突变
def mutation(population,pm):
# pm是概率阈值
px=len(population)
# 求出种群中所有种群/个体的个数
py=len(population[0])
# 染色体/个体基因的个数
for i in range(px):
if(random.random()<pm):
mpoint=random.randint(0,py-1)
# 生成0到py-1的随机数
if(population[i][mpoint]==1):
#将mpoint个基因进行单点随机变异，变为0或者1
population[i][mpoint]=0
else:
population[i][mpoint]=1
#transform the binary to decimalism
# 将每一个染色体都转化成十进制 max_value为基因最大值
def b2d(b,max_value,chromosome_length):
total=0
for i in range(len(b)):
total=total+b[i]*math.pow(2,i)
#从第一位开始，每一位对2求幂，然后求和，得到十进制数？
total=total*max_value/(math.pow(2,chromosome_length)-1)
return total
#寻找最好的适应度和个体
def best(population,fitness1):
px=len(population)
bestindividual=[]
bestfitness=fitness1[0]
for i in range(1,px):
# 循环找出最大的适应度，适应度最大的也就是最好的个体
if(fitness1[i]>bestfitness):
bestfitness=fitness1[i]
bestindividual=population[i]
return [bestindividual,bestfitness]
population_size=500
max_value=10
# 基因中允许出现的最大值
chromosome_length=10
pc=0.6
pm=0.01
results=[[]]
fitness1=[]
fitmean=[]
population=pop=species_origin(population_size,chromosome_length)
#生成一个初始的种群
for i in range(population_size):
function1=function(population,chromosome_length,max_value)
fitness1=fitness(function1)
best_individual,best_fitness=best(population,fitness1)
results.append([best_fitness,b2d(best_individual,max_value,chromosome_length)])
#将最好的个体和最好的适应度保存，并将最好的个体转成十进制
selection(population,fitness1)
crossover(population,pc)
mutation(population,pm)
results=results[1:]
results.sort()
X=[]
Y=[]
for i in range(500):
X.append(i)
Y.append(results[i][0])
plt.plot(X,Y)
plt.show()
# -*-coding:utf-8 -*-
#
