-
Notifications
You must be signed in to change notification settings - Fork 0
/
PSO.py
155 lines (125 loc) · 6.06 KB
/
PSO.py
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
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
#PARTICLE SWARM OPTIMIZATION
from __future__ import print_function
import numpy as np
import os
import time
def rosenbrock_fun(x):
"""Rosenbrock function
Domain: -5 < xi < 5
Global minimun: f_min(1,..,1)=0
a, b = 1, 100
f(x,y) = (a-x)^2 + b(y-x^2)^2
"""
return (1 - x[0])**2 + 100*(x[1] - x[0]**2)**2
def himmelblau_fun(x):
"""Himmelblau function
Domain: -4 < xi < 4
Global minimun: f_min(1,..,1)=0
f(x,y) = (x^2 + y - 11)^2 + (x - y^2 - 7)^2
"""
return (x[0]**2 + x[1] - 11)**2 + (x[0] + x[1]**2 - 7)**2
def pso_rose(func, bounds, swarm_size=10, inertia=0.5, pa=1.95, ga=2.05,
max_vnorm=10, num_iters=100, verbose=False, func_name=None):
bounds = np.array(bounds)
assert np.all(bounds[:,0] < bounds[:,1])
dim = len(bounds)
X = np.random.rand(swarm_size, dim)
print('## Optimize:',func_name)
def clip_by_norm(x, max_norm):
norm = np.linalg.norm(x)
return x if norm <=max_norm else x * max_norm / norm
particles = X * (bounds[:,1]-bounds[:,0]) + bounds[:,0]
velocities = X * (bounds[:,1]-bounds[:,0]) + bounds[:,0]
personal_bests = np.copy(particles)
personal_best_fitness = [np.inf for p in particles]
global_best_idx = np.argmin(personal_best_fitness)
global_best = personal_bests[global_best_idx]
global_best_fitness = func(global_best)
history = {'particles':[],
'global_best_fitness':[],
'global_best':[[np.inf, np.inf] for i in range(num_iters)],
'obj_func': func_name,}
for i in range(num_iters):
StartTime = time.time()
history['particles'].append(particles)
history['global_best_fitness'].append(global_best_fitness)
history['global_best'][i][0] = global_best[0]
history['global_best'][i][1] = global_best[1]
if verbose: print('iter# {}:'.format(i), end='')
for p_i in range(swarm_size):
fitness = func(particles[p_i])
if fitness < personal_best_fitness[p_i]:
personal_bests[p_i] = particles[p_i]
personal_best_fitness[p_i] = fitness
if np.min(personal_best_fitness) < global_best_fitness:
global_best_idx = np.argmin(personal_best_fitness)
global_best = personal_bests[global_best_idx]
global_best_fitness = func(global_best)
m = inertia * velocities
acc_local = pa * np.random.rand() * (personal_bests - particles)
acc_global = ga * np.random.rand() * (global_best - particles)
velocities = m + acc_local + acc_global
velocities = clip_by_norm(velocities, max_vnorm)
particles = particles + velocities
EndTime = time.time()
if verbose:
print(' Fitness:{:.6f}, Position:{}, Velocity:{}, Time:{:.9f}'.format(global_best_fitness, global_best, np.linalg.norm(velocities), EndTime - StartTime))
return history
def pso_himme(func, bounds, swarm_size=10, inertia=0.5, pa=1.95, ga=2.05,
max_vnorm=10, num_iters=100, verbose=False, func_name=None):
bounds = np.array(bounds)
assert np.all(bounds[:,0] < bounds[:,1])
dim = len(bounds)
X = np.random.rand(swarm_size, dim)
print('## Optimize:',func_name)
def clip_by_norm(x, max_norm):
norm = np.linalg.norm(x)
return x if norm <=max_norm else x * max_norm / norm
particles = X * (bounds[:,1]-bounds[:,0]) + bounds[:,0]
velocities = X * (bounds[:,1]-bounds[:,0]) + bounds[:,0]
personal_bests = np.copy(particles)
personal_best_fitness = [np.inf for p in particles]
global_best_idx = np.argmin(personal_best_fitness)
global_best = personal_bests[global_best_idx]
global_best_fitness = func(global_best)
history = {'particles':[],
'global_best_fitness':[],
'global_best':[[np.inf, np.inf] for i in range(num_iters)],
'obj_func': func_name,}
for i in range(num_iters):
StartTime = time.time()
history['particles'].append(particles)
history['global_best_fitness'].append(global_best_fitness)
history['global_best'][i][0] = global_best[0]
history['global_best'][i][1] = global_best[1]
if verbose: print('iter# {}:'.format(i), end='')
for p_i in range(swarm_size):
fitness = func(particles[p_i])
if fitness > personal_best_fitness[p_i]:
personal_bests[p_i] = particles[p_i]
personal_best_fitness[p_i] = fitness
if np.min(personal_best_fitness) > global_best_fitness:
global_best_idx = np.argmin(personal_best_fitness)
global_best = personal_bests[global_best_idx]
global_best_fitness = func(global_best)
m = inertia * velocities
acc_local = pa * np.random.rand() * (personal_bests - particles)
acc_global = ga * np.random.rand() * (global_best - particles)
velocities = m + acc_local + acc_global
velocities = clip_by_norm(velocities, max_vnorm)
particles = particles + velocities
EndTime = time.time()
if verbose:
print(' Fitness:{:.6f}, Position:{}, Velocity:{}, Time:{:.9f}'.format(global_best_fitness, global_best, np.linalg.norm(velocities), EndTime - StartTime))
return history
print("""
1) Función de rosenbrock
2) Funcion de Himmelblau
""")
op = int(input("\nIngrese su opción: "))
if op == 1:
history = pso_rose(himmelblau_fun, bounds=[[-4,4],[-4,4]], swarm_size=30, inertia=0.5, num_iters=50, verbose=1, func_name='Rosenbrock Function')
print('global best:',history['global_best'][-1], ', global best position:', history['global_best'][-1])
if op == 2:
history = pso_himme(rosenbrock_fun, bounds=[[-5,5],[-5,5]], swarm_size=30, inertia=0.5, num_iters=50, verbose=1, func_name='Himmelblau Function')
print('global best:',history['global_best_fitness'][-1], ', global best position:', history['global_best'][-1])