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
|
import pytest
import math
import dartpy as dart
# Problem reference: http://ab-initio.mit.edu/wiki/index.php/NLopt_Tutorial
class SampleObjFunc(dart.optimizer.Function):
def eval(self, x):
return math.sqrt(x[1])
def evalGradient(self, x, grad):
grad[0] = 0
grad[1] = 0.5 / (math.sqrt(x[1]) + 0.000001)
class SampleConstFunc(dart.optimizer.Function):
def __init__(self, a, b):
super(SampleConstFunc, self).__init__()
self.a = a
self.b = b
def eval(self, x):
return (self.a*x[0] + self.b) * (self.a*x[0] + self.b) * (self.a*x[0] + self.b) - x[1];
def evalGradient(self, x, grad):
grad[0] = 3 * self.a * (self.a*x[0] + self.b) * (self.a*x[0] + self.b);
grad[1] = -1.0;
def test_gradient_descent_solver():
prob = dart.optimizer.Problem(2)
assert prob.getDimension() is 2
prob.setLowerBounds([-1e100, 0])
prob.setInitialGuess([1.234, 5.678])
assert prob.getInitialGuess()[0] == pytest.approx(1.234)
obj = SampleObjFunc()
prob.setObjective(obj)
solver = dart.optimizer.GradientDescentSolver(prob)
success = solver.solve()
assert success is True
min_f = prob.getOptimumValue()
opt_x = prob.getOptimalSolution()
assert min_f == pytest.approx(0, 1e-6)
assert len(opt_x) == prob.getDimension()
assert opt_x[0] == pytest.approx(1.234, 0.0)
assert opt_x[1] == pytest.approx(0, solver.getTolerance())
def test_nlopt_solver():
if not hasattr(dart.optimizer, 'NloptSolver'):
return
prob = dart.optimizer.Problem(2)
assert prob.getDimension() is 2
prob.setLowerBounds([-1e100, 0])
prob.setInitialGuess([1.234, 5.678])
assert prob.getInitialGuess()[0] == pytest.approx(1.234)
obj = SampleObjFunc()
prob.setObjective(obj)
const1 = SampleConstFunc(2, 0)
const2 = SampleConstFunc(-1, 1)
prob.addIneqConstraint(const1)
prob.addIneqConstraint(const2)
solver = dart.optimizer.NloptSolver(prob)
solver.setAlgorithm(dart.optimizer.NloptSolver.Algorithm.LD_MMA)
success = solver.solve()
assert success is True
min_f = prob.getOptimumValue()
opt_x = prob.getOptimalSolution()
assert min_f == pytest.approx(0.544330847, 1e-5)
assert len(opt_x) == prob.getDimension()
assert opt_x[0] == pytest.approx(0.333334, 1e-5)
assert opt_x[1] == pytest.approx(0.296296, 1e-5)
if __name__ == "__main__":
pytest.main()
|
