|
1
|
|
|
import numpy as np |
|
2
|
|
|
import pandas as pd |
|
3
|
|
|
from gradient_free_optimizers import RandomSearchOptimizer |
|
4
|
|
|
|
|
5
|
|
|
|
|
6
|
|
|
def objective_function(para): |
|
7
|
|
|
score = -para["x1"] * para["x1"] |
|
8
|
|
|
return score |
|
9
|
|
|
|
|
10
|
|
|
|
|
11
|
|
|
search_space = { |
|
12
|
|
|
"x1": np.arange(0, 100000, 0.1), |
|
13
|
|
|
} |
|
14
|
|
|
|
|
15
|
|
|
|
|
16
|
|
|
def test_attributes_results_0(): |
|
17
|
|
|
opt = RandomSearchOptimizer(search_space) |
|
18
|
|
|
opt.search(objective_function, n_iter=100) |
|
19
|
|
|
|
|
20
|
|
|
assert isinstance(opt.results, pd.DataFrame) |
|
21
|
|
|
|
|
22
|
|
|
|
|
23
|
|
|
def test_attributes_results_1(): |
|
24
|
|
|
opt = RandomSearchOptimizer(search_space) |
|
25
|
|
|
opt.search(objective_function, n_iter=100) |
|
26
|
|
|
|
|
27
|
|
|
assert set(search_space.keys()) < set(opt.results.columns) |
|
28
|
|
|
|
|
29
|
|
|
|
|
30
|
|
|
def test_attributes_results_2(): |
|
31
|
|
|
opt = RandomSearchOptimizer(search_space) |
|
32
|
|
|
opt.search(objective_function, n_iter=100) |
|
33
|
|
|
|
|
34
|
|
|
assert "x1" in list(opt.results.columns) |
|
35
|
|
|
|
|
36
|
|
|
|
|
37
|
|
|
def test_attributes_results_3(): |
|
38
|
|
|
opt = RandomSearchOptimizer(search_space) |
|
39
|
|
|
opt.search(objective_function, n_iter=100) |
|
40
|
|
|
|
|
41
|
|
|
assert "score" in list(opt.results.columns) |
|
42
|
|
|
|
|
43
|
|
|
|
|
44
|
|
|
def test_attributes_results_4(): |
|
45
|
|
|
opt = RandomSearchOptimizer(search_space) |
|
46
|
|
|
opt.search( |
|
47
|
|
|
objective_function, n_iter=1, initialize={}, warm_start=[{"x1": 0}] |
|
48
|
|
|
) |
|
49
|
|
|
|
|
50
|
|
|
assert 0 in list(opt.results["x1"].values) |
|
51
|
|
|
|
|
52
|
|
|
|
|
53
|
|
|
def test_attributes_results_5(): |
|
54
|
|
|
opt = RandomSearchOptimizer(search_space) |
|
55
|
|
|
opt.search( |
|
56
|
|
|
objective_function, n_iter=1, initialize={}, warm_start=[{"x1": 10}] |
|
57
|
|
|
) |
|
58
|
|
|
|
|
59
|
|
|
assert 10 in list(opt.results["x1"].values) |
|
60
|
|
|
|
|
61
|
|
|
|
|
62
|
|
|
def test_attributes_results_6(): |
|
63
|
|
|
def objective_function(para): |
|
64
|
|
|
score = -para["x1"] * para["x1"] |
|
65
|
|
|
return score |
|
66
|
|
|
|
|
67
|
|
|
search_space = { |
|
68
|
|
|
"x1": np.arange(0, 10, 1), |
|
69
|
|
|
} |
|
70
|
|
|
|
|
71
|
|
|
opt = RandomSearchOptimizer(search_space) |
|
72
|
|
|
opt.search( |
|
73
|
|
|
objective_function, n_iter=20, initialize={"random": 1}, memory=False |
|
74
|
|
|
) |
|
75
|
|
|
|
|
76
|
|
|
x1_results = list(opt.results["x1"].values) |
|
77
|
|
|
|
|
78
|
|
|
print("\n x1_results \n", x1_results) |
|
79
|
|
|
|
|
80
|
|
|
assert len(set(x1_results)) < len(x1_results) |
|
81
|
|
|
|
|
82
|
|
|
|
|
83
|
|
|
""" |
|
84
|
|
|
def test_attributes_results_7(): |
|
85
|
|
|
def objective_function(para): |
|
86
|
|
|
score = -para["x1"] * para["x1"] |
|
87
|
|
|
return score |
|
88
|
|
|
|
|
89
|
|
|
search_space = { |
|
90
|
|
|
"x1": np.arange(0, 10, 1), |
|
91
|
|
|
} |
|
92
|
|
|
|
|
93
|
|
|
opt = RandomSearchOptimizer(search_space) |
|
94
|
|
|
opt.search( |
|
95
|
|
|
objective_function, n_iter=20, initialize={"random": 1}, memory=True |
|
96
|
|
|
) |
|
97
|
|
|
|
|
98
|
|
|
x1_results = list(opt.results["x1"].values) |
|
99
|
|
|
|
|
100
|
|
|
print("\n x1_results \n", x1_results) |
|
101
|
|
|
|
|
102
|
|
|
assert len(set(x1_results)) == len(x1_results) |
|
103
|
|
|
|
|
104
|
|
|
|
|
105
|
|
|
def test_attributes_results_8(): |
|
106
|
|
|
def objective_function(para): |
|
107
|
|
|
score = -para["x1"] * para["x1"] |
|
108
|
|
|
return score |
|
109
|
|
|
|
|
110
|
|
|
search_space = { |
|
111
|
|
|
"x1": np.arange(-10, 11, 1), |
|
112
|
|
|
} |
|
113
|
|
|
|
|
114
|
|
|
results = pd.DataFrame(np.arange(-10, 10, 1), columns=["x1"]) |
|
115
|
|
|
results["score"] = 0 |
|
116
|
|
|
|
|
117
|
|
|
opt = RandomSearchOptimizer(search_space) |
|
118
|
|
|
opt.search( |
|
119
|
|
|
objective_function, |
|
120
|
|
|
n_iter=100, |
|
121
|
|
|
initialize={}, |
|
122
|
|
|
memory=True, |
|
123
|
|
|
memory_warm_start=results, |
|
124
|
|
|
) |
|
125
|
|
|
|
|
126
|
|
|
print("\n opt.results \n", opt.results) |
|
127
|
|
|
|
|
128
|
|
|
x1_results = list(opt.results["x1"].values) |
|
129
|
|
|
|
|
130
|
|
|
assert 10 == x1_results[0] |
|
131
|
|
|
""" |
|
132
|
|
|
|
SimonBlanke /
Gradient-Free-Optimizers
