作者：邬杨明
1 import numpy
2 import matplotlib.pyplot as plt
3 import random
4 import operator
5
6 x_label=[]
7 y_label=[]#将每一步迭代的结果存储到列表中,便于画图
8 class GA(object):
9 def __init__(self,length,number,lower_boundary,upper_boundary,iter_number):
10 self.length = length#确定染色体编码长度
11 self.number = number#确定初始化种群数量
12 self.lower_boundary = lower_boundary#定义域下界
13 self.upper_boundary = upper_boundary#定义域上届
14 self.population = self.initial_population(length,number)
15 self.iteration(iter_number)
16 def initial_population(self,length,number):#初始化种群
17 return [self.initial_chromosome(length) for i in range(number)]
18 def initial_chromosome(self,length):#编码
19 """
20 随机生成长度为length的染色体，每个基因的取值是0或1
21 这里用一个bit表示一个基因
22 """
23 chromosome = 0
24 for i in range(self.length):
25 chromosome |= (1 << i) * random.randint(0, 1)
26 # print(chromosome)
27 # print(num)
28 return chromosome
29 def decode(self,chromosome):#解码
30 x = chromosome*(self.upper_boundary-self.lower_boundary)/(2**self.length-1)
31 return x
32 def fitness_function(self,chromosome):#适应度函数
33 x = self.decode(chromosome)
34 y = x + 10 * numpy.sin(5 * x) + 7 * numpy.cos(4 * x)
35 return y
36 def evolution(self, retain_rate=0.2, random_select_rate=0.5, mutation_rate=0.01):
37 """
38 对当前种群依次进行选择、交叉并生成新一代种群，然后对新一代种群进行变异
39 """
40 parents = self.selection(retain_rate, random_select_rate)
41 self.crossover(parents)
42 self.mutation(mutation_rate)
43 def selection(self,retain_rate, random_select_rate):
44 graded = [(self.fitness_function(chromosome), chromosome) for chromosome in self.population]
45 sort = [x[1] for x in sorted(graded, reverse=True)]
46 retain_length=int(len(sort)*retain_rate)
47 #选出适应性强的个体，精英选择
48 parents=sort[:retain_length]
49 for chromosome in sort[retain_length:]:
50 if random.random() < random_select_rate:
51 parents.append(chromosome)
52 return parents
53 def crossover(self,parents):#交叉
54 children=[]
55 #需要繁殖的子代数量
56 target_number=len(self.population)-len(parents)
57 #开始繁殖
58 while len(children) < target_number:
59 father = random.randint(0, len(parents) - 1)
60 mother = random.randint(0, len(parents) - 1)
61 if father != mother:
62 # 随机选取交叉点
63 cross_point = random.randint(0, self.length)
64 # 生成掩码，方便位操作
65 mark = 0
66 for i in range(cross_point):
67 mark |= (1 << i)
68 father = parents[father]
69 mother = parents[mother]
70 # 孩子将获得父亲在交叉点前的基因和母亲在交叉点后（包括交叉点）的基因
71 child = ((father & mark) | (mother & ~mark)) & ((1 << self.length) - 1)
72 children.append(child)
73 # 经过繁殖后，孩子和父母的数量与原始种群数量相等，在这里可以更新种群。
74 self.population = parents + children
75 def mutation(self,rate):
76 for i in range(len(self.population)):
77 if random.random() < rate:
78 j = random.randint(0, self.length - 1)#s随机产生变异位置
79 self.population[i] ^= 1 << j #产生变异
80 def result(self):
81 '''
82 获得近优值
83 '''
84 graded = [(self.fitness_function(chromosome), chromosome) for chromosome in self.population]
85 graded = [x[1] for x in sorted(graded, reverse=True)]
86 # print('近优的x值为：', self.decode(graded[0]),
87 # '近优的y值为：', self.fitness_function(graded[0]))
88 #将每一步迭代的结果存储到列表中
89 y_label.append(self.fitness_function(graded[0]))
90 return self.decode(graded[0])
91 def iteration(self,iter_number):
92 for i in range(iter_number):
93 self.evolution()
94 self.result()
95 x_label.append(i)
96 g1=GA(17,300,0,9,30)#建立一个实例
97
98 plt.plot(x_label,y_label)
99 plt.show()
100 g1.result()
