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
|
#!/usr/bin/env python
from __future__ import division
import time, sys, os, numpy
from cogent.util.unit_test import TestCase, main
from cogent.maths.optimisers import maximise, MaximumEvaluationsReached
__author__ = "Peter Maxwell and Gavin Huttley"
__copyright__ = "Copyright 2007-2012, The Cogent Project"
__credits__ = ["Peter Maxwell", "Gavin Huttley"]
__license__ = "GPL"
__version__ = "1.5.3"
__maintainer__ = "Gavin Huttley"
__email__ = "gavin.huttley@anu.edu.au"
__status__ = "Production"
def quartic(x):
# Has global maximum at -4 and local maximum at 2
# http://www.wolframalpha.com/input/?i=x**2*%283*x**2%2B8*x-48%29
# Scaled down 10-fold to avoid having to change init_temp
return x**2*(3*x**2+8*x-48)
class NullFile(object):
def write(self, x):
pass
def isatty(self):
return False
def quiet(f, *args, **kw):
# Checkpointer still has print statements
orig = sys.stdout
try:
sys.stdout = NullFile()
result = f(*args, **kw)
finally:
sys.stdout = orig
return result
def MakeF():
evals = [0]
last = [0]
def f(x):
evals[0] += 1
last[0] = x
# Scaled down 10-fold to avoid having to change init_temp
return -0.1 * quartic(x)
return f, last, evals
class OptimiserTestCase(TestCase):
def _test_optimisation(self, target=-4, xinit=1.0, bounds=([-10,10]), **kw):
local = kw.get('local', None)
max_evaluations = kw.get('max_evaluations', None)
f, last, evals = MakeF()
x = quiet(maximise, f, [xinit], bounds, **kw)
self.assertEqual(x, last[0]) # important for Calculator
error = abs(x[0] - target)
self.assertTrue(error < .0001, (kw, x, target, x))
def test_global(self):
# Should find global minimum
self._test_optimisation(local=False, seed=1)
def test_bounded(self):
# Global minimum out of bounds, so find secondary one
# numpy.seterr('raise')
self._test_optimisation(bounds=([0.0],[10.0]), target=2, seed=1)
def test_local(self):
# Global minimum not the nearest one
self._test_optimisation(local=True, target=2)
def test_limited(self):
self.assertRaises(MaximumEvaluationsReached,
self._test_optimisation, max_evaluations=5)
# def test_limited_warning(self):
# """optimiser warning if max_evaluations exceeded"""
# self._test_optimisation(max_evaluations=5, limit_action='warn')
def test_get_max_eval_count(self):
"""return the evaluation count from optimisation"""
f, last, evals = MakeF()
x, e = quiet(maximise, f, xinit=[1.0], bounds=([-10,10]),
return_eval_count=True)
self.assertTrue(e > 500)
def test_checkpointing(self):
filename = 'checkpoint.tmp.pickle'
if os.path.exists(filename):
os.remove(filename)
self._test_optimisation(filename=filename, seed=1, init_temp=10)
self._test_optimisation(filename=filename, seed=1, init_temp=10)
self.assertRaises(Exception, self._test_optimisation,
filename=filename, seed=1, init_temp=3.21)
if os.path.exists(filename):
os.remove(filename)
if __name__ == '__main__':
main()
|
