-
Notifications
You must be signed in to change notification settings - Fork 0
Expand file tree
/
Copy pathpsoClass.py
More file actions
97 lines (94 loc) · 3.65 KB
/
Copy pathpsoClass.py
File metadata and controls
97 lines (94 loc) · 3.65 KB
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
# -*- coding:utf-8 -*-
import numpy as np
import random
class PSO:
def __init__(self):
# 参数初始化
self.c1 = 0.4 # 粒子自身过去的位置对粒子迁移速度的影响
self.c2 = 1.4 # 粒子群中最好的粒子的位置对粒子迁移速度的影响
self.maxgen = 200 # 进化代数
self.popsize = 100 # 粒子群大小
self.psize = 3 # 粒子大小--> 指的是维度
self.Vmax = 5
self.Vmin = -5 # 粒子移动速度
self.lb = np.ones((self.psize,1))*(-10)
self.ub = np.ones((self.psize,1))*10
self.bound = np.hstack((self.lb,self.ub)) # 粒子移动范围
self.wstart = 0.9;self.wend = 0.4 # 惯性权重
# 返回0~1二维数组
def rand(self,r,c):
mat = np.zeros((r,c)) # 创建初始r*c数组
for i in range(r):
for j in range(c):
mat[i][j] = random.random()
return mat
# 返回一个二维数组的最小值和索引
def amin_index(self,a,axis=0):
"""
a是一个矩阵
"""
value_min = np.amin(a,axis=axis)
index_min = []
if axis == 0:
for i in range(a.shape[1]):#以矩阵a的列长度作为循环条件
index_min.append(list(a[:,i]).index(value_min[i]))
else:
for i in range(a.shape[0]):
index_min.append(list(a[i,:]).index(value_min[i]))
index_min = np.array(index_min)
return value_min,index_min
# 返回一个二维数组的最大值和索引
def amax_index(self,a,axis=0):
value_max = np.amax(a,axis=axis)
index_max = []
if axis == 0:
for i in range(a.shape[1]):
index_max.append(list(a[:,i]).index(value_max[i]))
else:
for i in range(a.shape[0]):
index_max.append(list(a[i,:]).index(value_max[i]))
index_min = np.array(index_max)
return value_max,index_min
# 适应度函数,x为一维数组
def fun(self,x):
x1 = x[0]
x2 = x[1]
x3 = x[2]
fitness = x1**2+x2**2-x3**2+3*np.sin(x1*x2*x3)
fitness = -1.0*fitness # 求函数最小值时注释掉这行代码
return fitness
# 种群初始化
def initial(self):
# 初始化种群
group = {}
group['pop'] = np.zeros((self.popsize,self.psize)) # 种群
group['v'] = np.zeros((self.popsize,self.psize)) # 粒子速度
group['fitness'] = np.zeros(self.popsize) # 粒子适应度值，一维数组
for i in range(self.popsize):
group['pop'][i] = self.bound[:,0]+(self.bound[:,1]-self.bound[:,0])*self.rand(1,self.psize)[0]
group['v'][i] = np.ones(3)*self.Vmin+(np.ones(3)*self.Vmax-np.ones(3)*self.Vmin)*self.rand(1,3)[0]
group['fitness'][i] = self.fun(group['pop'][i])
return group
# 限制粒子的速度范围
def f_av(self,av): # 对向量av进行边界限制
def f_v(v): # 对单一的v进行边界限制
if v > self.Vmax:
return self.Vmax
elif v < self.Vmin:
return self.Vmin
else:
return v
av = np.array(map(f_v,av))
return av
# 限制粒子的速度范围
def f_ap(self,ap): # 对向量ap进行边界限制
def f_p(p,p_index): # 对单一的p进行边界限制
if p > self.bound[p_index,1]:
return self.bound[p_index,1]
elif p < self.bound[p_index,0]:
return self.bound[p_index,0]
else:
return p
for i in range(len(ap)):
ap[i] = f_p(ap[i],i)
return ap