|
1
|
|
|
# encoding=utf8 |
|
2
|
|
|
import copy |
|
3
|
|
|
import logging |
|
4
|
|
|
|
|
5
|
|
|
from numpy import asarray, full, argmax |
|
6
|
|
|
|
|
7
|
|
|
from NiaPy.algorithms.algorithm import Algorithm, Individual, defaultIndividualInit |
|
8
|
|
|
|
|
9
|
|
|
logging.basicConfig() |
|
10
|
|
|
logger = logging.getLogger('NiaPy.algorithms.basic') |
|
11
|
|
|
logger.setLevel('INFO') |
|
12
|
|
|
|
|
13
|
|
|
__all__ = ['ArtificialBeeColonyAlgorithm'] |
|
14
|
|
|
|
|
15
|
|
|
class SolutionABC(Individual): |
|
16
|
|
|
r"""Representation of solution for Artificial Bee Colony Algorithm. |
|
17
|
|
|
|
|
18
|
|
|
Date: |
|
19
|
|
|
2018 |
|
20
|
|
|
|
|
21
|
|
|
Author: |
|
22
|
|
|
Klemen Berkovič |
|
23
|
|
|
|
|
24
|
|
|
See Also: |
|
25
|
|
|
* :class:`NiaPy.algorithms.Individual` |
|
26
|
|
|
""" |
|
27
|
|
|
def __init__(self, **kargs): |
|
28
|
|
|
r"""Initialize individual. |
|
29
|
|
|
|
|
30
|
|
|
Args: |
|
31
|
|
|
kargs (Dict[str, Any]): Additional arguments. |
|
32
|
|
|
|
|
33
|
|
|
See Also: |
|
34
|
|
|
* :func:`NiaPy.algorithms.Individual.__init__` |
|
35
|
|
|
""" |
|
36
|
|
|
Individual.__init__(self, **kargs) |
|
37
|
|
|
|
|
38
|
|
|
class ArtificialBeeColonyAlgorithm(Algorithm): |
|
39
|
|
|
r"""Implementation of Artificial Bee Colony algorithm. |
|
40
|
|
|
|
|
41
|
|
|
Algorithm: |
|
42
|
|
|
Artificial Bee Colony algorithm |
|
43
|
|
|
|
|
44
|
|
|
Date: |
|
45
|
|
|
2018 |
|
46
|
|
|
|
|
47
|
|
|
Author: |
|
48
|
|
|
Uros Mlakar and Klemen Berkovič |
|
49
|
|
|
|
|
50
|
|
|
License: |
|
51
|
|
|
MIT |
|
52
|
|
|
|
|
53
|
|
|
Reference paper: |
|
54
|
|
|
Karaboga, D., and Bahriye B. "A powerful and efficient algorithm for numerical function optimization: artificial bee colony (ABC) algorithm." Journal of global optimization 39.3 (2007): 459-471. |
|
55
|
|
|
|
|
56
|
|
|
Arguments |
|
57
|
|
|
Name (List[str]): List containing strings that represent algorithm names |
|
58
|
|
|
Limit (Union[float, numpy.ndarray[float]]): Limt |
|
59
|
|
|
|
|
60
|
|
|
See Also: |
|
61
|
|
|
* :class:`NiaPy.algorithms.Algorithm` |
|
62
|
|
|
""" |
|
63
|
|
|
Name = ['ArtificialBeeColonyAlgorithm', 'ABC'] |
|
64
|
|
|
|
|
65
|
|
|
@staticmethod |
|
66
|
|
|
def typeParameters(): |
|
67
|
|
|
r"""Return functions for checking values of parameters. |
|
68
|
|
|
|
|
69
|
|
|
Returns: |
|
70
|
|
|
Dict[str, Callable]: |
|
71
|
|
|
* Limit (Callable[Union[float, numpy.ndarray[float]]]): TODO |
|
72
|
|
|
|
|
73
|
|
|
See Also: |
|
74
|
|
|
* :func:`NiaPy.algorithms.Algorithm.typeParameters` |
|
75
|
|
|
""" |
|
76
|
|
|
d = Algorithm.typeParameters() |
|
77
|
|
|
d.update({'Limit': lambda x: isinstance(x, int) and x > 0}) |
|
78
|
|
|
return d |
|
79
|
|
|
|
|
80
|
|
|
def setParameters(self, NP=10, Limit=100, **ukwargs): |
|
81
|
|
|
r"""Set the parameters of Artificial Bee Colony Algorithm. |
|
82
|
|
|
|
|
83
|
|
|
Parameters: |
|
84
|
|
|
Limit (Optional[Union[float, numpy.ndarray[float]]]): Limt |
|
85
|
|
|
**ukwargs (Dict[str, Any]): Additional arguments |
|
86
|
|
|
|
|
87
|
|
|
See Also: |
|
88
|
|
|
* :func:`NiaPy.algorithms.Algorithm.setParameters` |
|
89
|
|
|
""" |
|
90
|
|
|
Algorithm.setParameters(self, NP=NP, InitPopFunc=defaultIndividualInit, itype=SolutionABC, **ukwargs) |
|
91
|
|
|
self.FoodNumber, self.Limit = int(self.NP / 2), Limit |
|
92
|
|
|
|
|
93
|
|
|
def CalculateProbs(self, Foods, Probs): |
|
94
|
|
|
r"""Calculate the probes. |
|
95
|
|
|
|
|
96
|
|
|
Parameters: |
|
97
|
|
|
Foods (numpy.ndarray): TODO |
|
98
|
|
|
Probs (numpy.ndarray): TODO |
|
99
|
|
|
|
|
100
|
|
|
Returns: |
|
101
|
|
|
numpy.ndarray: TODO |
|
102
|
|
|
""" |
|
103
|
|
|
Probs = [1.0 / (Foods[i].f + 0.01) for i in range(self.FoodNumber)] |
|
104
|
|
|
s = sum(Probs) |
|
105
|
|
|
Probs = [Probs[i] / s for i in range(self.FoodNumber)] |
|
106
|
|
|
return Probs |
|
107
|
|
|
|
|
108
|
|
|
def initPopulation(self, task): |
|
109
|
|
|
r"""Initialize the starting population. |
|
110
|
|
|
|
|
111
|
|
|
Parameters: |
|
112
|
|
|
task (Task): Optimization task |
|
113
|
|
|
|
|
114
|
|
|
Returns: |
|
115
|
|
|
Tuple[numpy.ndarray, numpy.ndarray[float], Dict[str, Any]]: |
|
116
|
|
|
1. New population |
|
117
|
|
|
2. New population fitness/function values |
|
118
|
|
|
3. Additional arguments: |
|
119
|
|
|
* Probes (numpy.ndarray): TODO |
|
120
|
|
|
* Trial (numpy.ndarray): TODO |
|
121
|
|
|
|
|
122
|
|
|
See Also: |
|
123
|
|
|
* :func:`NiaPy.algorithms.Algorithm.initPopulation` |
|
124
|
|
|
""" |
|
125
|
|
|
Foods, fpop, _ = Algorithm.initPopulation(self, task) |
|
126
|
|
|
Probs, Trial = full(self.FoodNumber, 0.0), full(self.FoodNumber, 0.0) |
|
127
|
|
|
return Foods, fpop, {'Probs': Probs, 'Trial': Trial} |
|
128
|
|
|
|
|
129
|
|
|
def runIteration(self, task, Foods, fpop, xb, fxb, Probs, Trial, **dparams): |
|
130
|
|
|
r"""Core funciton of the algorithm. |
|
131
|
|
|
|
|
132
|
|
|
Parameters: |
|
133
|
|
|
task (Task): Optimization task |
|
134
|
|
|
Foods (numpy.ndarray): Current population |
|
135
|
|
|
fpop (numpy.ndarray[float]): Function/fitness values of current population |
|
136
|
|
|
xb (numpy.ndarray): Current best individual |
|
137
|
|
|
fxb (float): Current best individual fitness/function value |
|
138
|
|
|
Probs (numpy.ndarray): TODO |
|
139
|
|
|
Trial (numpy.ndarray): TODO |
|
140
|
|
|
dparams (Dict[str, Any]): Additional parameters |
|
141
|
|
|
|
|
142
|
|
|
Returns: |
|
143
|
|
|
Tuple[numpy.ndarray, numpy.ndarray, numpy.ndarray, float, Dict[str, Any]]: |
|
144
|
|
|
1. New population |
|
145
|
|
|
2. New population fitness/function values |
|
146
|
|
|
3. New global best solution |
|
147
|
|
|
4. New global best fitness/objecive value |
|
148
|
|
|
5. Additional arguments: |
|
149
|
|
|
* Probes (numpy.ndarray): TODO |
|
150
|
|
|
* Trial (numpy.ndarray): TODO |
|
151
|
|
|
""" |
|
152
|
|
|
for i in range(self.FoodNumber): |
|
153
|
|
|
newSolution = copy.deepcopy(Foods[i]) |
|
154
|
|
|
param2change = int(self.rand() * task.D) |
|
155
|
|
|
neighbor = int(self.FoodNumber * self.rand()) |
|
156
|
|
|
newSolution.x[param2change] = Foods[i].x[param2change] + (-1 + 2 * self.rand()) * (Foods[i].x[param2change] - Foods[neighbor].x[param2change]) |
|
157
|
|
|
newSolution.evaluate(task, rnd=self.Rand) |
|
158
|
|
|
if newSolution.f < Foods[i].f: |
|
159
|
|
|
Foods[i], Trial[i] = newSolution, 0 |
|
160
|
|
|
if newSolution.f < fxb: xb, fxb = newSolution.x.copy(), newSolution.f |
|
161
|
|
|
else: Trial[i] += 1 |
|
162
|
|
|
Probs, t, s = self.CalculateProbs(Foods, Probs), 0, 0 |
|
163
|
|
|
while t < self.FoodNumber: |
|
164
|
|
|
if self.rand() < Probs[s]: |
|
165
|
|
|
t += 1 |
|
166
|
|
|
Solution = copy.deepcopy(Foods[s]) |
|
167
|
|
|
param2change = int(self.rand() * task.D) |
|
168
|
|
|
neighbor = int(self.FoodNumber * self.rand()) |
|
169
|
|
|
while neighbor == s: neighbor = int(self.FoodNumber * self.rand()) |
|
170
|
|
|
Solution.x[param2change] = Foods[s].x[param2change] + (-1 + 2 * self.rand()) * (Foods[s].x[param2change] - Foods[neighbor].x[param2change]) |
|
171
|
|
|
Solution.evaluate(task, rnd=self.Rand) |
|
172
|
|
|
if Solution.f < Foods[s].f: |
|
173
|
|
|
Foods[s], Trial[s] = Solution, 0 |
|
174
|
|
|
if Solution.f < fxb: xb, fxb = Solution.x.copy(), Solution.f |
|
175
|
|
|
else: Trial[s] += 1 |
|
176
|
|
|
s += 1 |
|
177
|
|
|
if s == self.FoodNumber: s = 0 |
|
178
|
|
|
mi = argmax(Trial) |
|
179
|
|
|
if Trial[mi] >= self.Limit: |
|
180
|
|
|
Foods[mi], Trial[mi] = SolutionABC(task=task, rnd=self.Rand), 0 |
|
181
|
|
|
if Foods[mi].f < fxb: xb, fxb = Foods[mi].x.copy(), Foods[mi].f |
|
182
|
|
|
return Foods, asarray([f.f for f in Foods]), xb, fxb, {'Probs': Probs, 'Trial': Trial} |
|
183
|
|
|
|
|
184
|
|
|
# vim: tabstop=3 noexpandtab shiftwidth=3 softtabstop=3 |
|
185
|
|
|
|
NiaOrg /
NiaPy
