-
Notifications
You must be signed in to change notification settings - Fork 1
/
krill_herd.py
216 lines (134 loc) · 5.39 KB
/
krill_herd.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
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
import numpy as np
import random
import math
import time
def print_krills(krills):
for i in krills:
print(i.X)
def l2_norm(X):
return np.linalg.norm(X)
res = []
class Krill:
def __init__(self, n, obj_func, ub, lb):
self.X = lb + [random.random() for i in range(n)] * (ub - lb)
self.N = np.zeros((n,))
self.F = np.zeros((n,))
self.X_best = self.X[:]
self.K_best = obj_func(self.X_best)
self.K = self.K_best
class KrillHerd():
def pos_effect(self, X_i, X_j):
return (X_j - X_i) / (l2_norm(X_j - X_i) + self.eps)
def fitness_effect(self, K_i, K_j):
return (K_i - K_j) / (self.K_ibest - self.K_iworst + self.eps)
def collect_neighbors(self, krill_i):
d_s = 0
for j in range(self.nk):
d_s += l2_norm(self.krills[j].X - krill_i.X)
d_s /= (5 * self.nk)
neighbors = list()
for j in range(self.nk):
if l2_norm(self.krills[j].X - krill_i.X) <= d_s:
neighbors.append(self.krills[j])
return neighbors
def neighbors_motion(self, krill_i):
neighbors = self.collect_neighbors(krill_i)
alpha_loc = 0
for i in neighbors:
alpha_loc += self.fitness_effect(krill_i.K, i.K) * self.pos_effect(krill_i.X, i.X)
return alpha_loc
def target_motion(self, krill_i, it):
return 2 * (random.random() + it/self.iter) * self.fitness_effect(krill_i.K, self.K_ibest) * self.pos_effect(krill_i.X, self.X_ibest)
def foraging_motion(self, krill_i, it):
X_food = 0
tmp = 0
for i in range(self.nk):
X_food += (self.krills[i].X/self.krills[i].K)
tmp += (1/self.krills[i].K)
X_food /= tmp
return 2 * (1 - it/self.iter) * self.pos_effect(krill_i.X, X_food) * self.fitness_effect(krill_i.K , self.obj_func(X_food))
def target_foraging(self, krill_i):
return self.fitness_effect(krill_i.K, krill_i.K_best) * self.pos_effect(krill_i.X, krill_i.X_best)
def __init__(self, n_dim = 1, obj_func = None, wn = 0.42, V_f = 0.02, wf = 0.38, eps = 1e-30, iter = 100, nk = 50, N_max=0.01, D_max=0.005, ub=1e10, lb=-1e10):
global res
self.n_dim, self.obj_func, self.wn, self.V_f, self.wf, self.eps, self.iter, self.nk, self.N_max, self.D_max = n_dim, obj_func, wn, V_f, wf, eps, iter, nk, N_max, D_max
self.ub = ub * np.ones((n_dim, ))
self.lb = lb * np.ones((n_dim, ))
self.krills = list()
for i in range(nk):
self.krills.append(Krill(n_dim, obj_func, self.ub, self.lb))
# print_krills(self.krills)
self.K_ibest = self.krills[0].K_best
self.K_iworst = self.K_ibest
self.X_ibest = list(self.krills[0].X_best)
for i in range(nk - 1):
if (self.krills[i + 1].K_best < self.K_ibest):
self.K_ibest = self.krills[i + 1].K_best
self.X_ibest = list(self.krills[i + 1].X_best)
self.K_iworst = max(self.K_iworst, self.krills[i + 1].K_best)
for it in range(self.iter):
for i in range(nk):
# Other krills induced motion
alpha_loc = self.neighbors_motion(self.krills[i])
alpha_target = self.target_motion(self.krills[i], it)
self.krills[i].N = alpha_loc * self.N_max + alpha_target * self.krills[i].N
# Foraging Motion
beta_food = self.foraging_motion(self.krills[i], it)
beta_best = self.target_foraging(self.krills[i])
self.krills[i].F = self.V_f * (beta_best + beta_food) + self.wf * self.krills[i].F
# Physical Diffusion
delta = ([2 * random.random() for k in range(self.n_dim)]) - np.ones((self.n_dim, ))
D_i = self.D_max * (1 - it/self.iter) * delta
# Time scaling factor delta_t
delta_t = 1.5 * np.sum(ub - lb)
# Updates
self.krills[i].X += delta_t * (self.krills[i].N + self.krills[i].F + D_i)
for x in range(self.n_dim):
if (self.krills[i].X[x] > self.ub[x]):
self.krills[i].X[x] = self.lb[x] + (self.ub[x] - self.lb[x]) * random.random()
if (self.krills[i].X[x] < self.lb[x]):
self.krills[i].X[x] = self.lb[x] + (self.ub[x] - self.lb[x]) * random.random()
# self.krills[i].X[x] = min(self.krills[i].X[x], self.ub[0])
# self.krills[i].X[x] = max(self.krills[i].X[x], self.lb[0])
self.krills[i].K = self.obj_func(self.krills[i].X)
if (self.krills[i].K < self.krills[i].K_best):
self.krills[i].K_best = self.krills[i].K
self.krills[i].X_best = self.krills[i].X[:]
# if (self.krills[i].K < self.K_ibest):
# self.K_ibest = self.krills[i].K
# self.X_ibest = list(self.krills[i].X)
# res = i
# print("HERD", i, self.krills[i].X, res, self.K_ibest)
# elif (self.krills[i].K > self.K_iworst):
# self.K_iworst = self.krills[i].K
for i in range(nk - 1):
if (self.krills[i + 1].K_best < self.K_ibest):
self.K_ibest = self.krills[i + 1].K_best
self.X_ibest = list(self.krills[i + 1].X_best)
print("HERD", i, self.krills[i + 1].X_best, res, self.K_ibest)
self.K_iworst = max(self.K_iworst, self.krills[i + 1].K_best)
# if i == 0:
# print("KRILL", i)
# print("X", self.krills[i].X)
# print("K", self.krills[i].K)
# print("N", self.krills[i].N)
# print("F", self.krills[i].F)
# print("D", D_i)
# time.sleep(5)
print("FINAL", self.X_ibest, self.K_ibest)
# time.sleep(5)
def peak(X):
return X[0] * math.exp(-(X[0]**2 + X[1]**2))
def sphere(X):
res = 0
for i in X:
res += i ** 2
return res
def main():
KrillHerd(n_dim=3, obj_func=sphere, ub=10, lb=-10, iter=100, nk=50)
# print(res)
# X = [0.17248473654989216, -0.3643013151473621, -1.883153284994198, -2.5369518187105893, -2.4602427456130727]
# print(X)
# print(sphere(X))
if __name__ == "__main__":
main()