1 /*
2 * Licensed to the Hipparchus project under one or more
3 * contributor license agreements. See the NOTICE file distributed with
4 * this work for additional information regarding copyright ownership.
5 * The Hipparchus project licenses this file to You under the Apache License, Version 2.0
6 * (the "License"); you may not use this file except in compliance with
7 * the License. You may obtain a copy of the License at
8 *
9 * https://www.apache.org/licenses/LICENSE-2.0
10 *
11 * Unless required by applicable law or agreed to in writing, software
12 * distributed under the License is distributed on an "AS IS" BASIS,
13 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
14 * See the License for the specific language governing permissions and
15 * limitations under the License.
16 */
17 package org.hipparchus.optim.nonlinear.vector.constrained;
18
19 import org.junit.Assert;
20 import org.hipparchus.linear.MatrixUtils;
21 import org.hipparchus.linear.RealVector;
22 import org.hipparchus.optim.InitialGuess;
23 import org.hipparchus.optim.OptimizationData;
24 import org.hipparchus.optim.nonlinear.scalar.ObjectiveFunction;
25
26 public abstract class AbstractConstrainedOptimizerTest {
27
28 /** Build the optimizer.
29 * @return built optimizer
30 */
31 protected abstract ConstraintOptimizer buildOptimizer();
32
33 /** Test one problem.
34 * @param expectedSolution expected solution
35 * @param solutionTolerance tolerance on solution (L₁ norm)
36 * @param expectedMultipliers expected multipliers
37 * @param multipliersTolerance tolerance on multipliers (L₁ norm)
38 * @param expectedValue expected objective function value
39 * @param valueTolerance tolerance on objective function value
40 * @param objectiveFunction objective function
41 * @param initialGuess initial guess (may be null)
42 * @param constraints contraints
43 */
44 protected void doTestProblem(final double[] expectedSolution,
45 final double solutionTolerance,
46 final double[] expectedMultipliers,
47 final double multipliersTolerance,
48 final double expectedValue,
49 final double valueTolerance,
50 final ObjectiveFunction objectiveFunction,
51 final double[] initialGuess,
52 final Constraint... constraints) {
53
54 // find optimum solution
55 final ConstraintOptimizer optimizer = buildOptimizer();
56 final OptimizationData[] data = new OptimizationData[constraints.length + (initialGuess == null ? 1 : 2)];
57 data[0] = objectiveFunction;
58 System.arraycopy(constraints, 0, data, 1, constraints.length);
59 if (initialGuess != null) {
60 data[data.length - 1] = new InitialGuess(initialGuess);
61 }
62 final LagrangeSolution solution = optimizer.optimize(data);
63
64 // check result
65 Assert.assertEquals(0.0,
66 MatrixUtils.createRealVector(expectedSolution).subtract(solution.getX()).getL1Norm(),
67 solutionTolerance);
68 Assert.assertEquals(0.0,
69 MatrixUtils.createRealVector(expectedMultipliers).subtract(solution.getLambda()).getL1Norm(),
70 multipliersTolerance);
71 Assert.assertEquals(expectedValue, solution.getValue(), valueTolerance);
72
73 // check neighboring points either violate constraints or have worst objective function
74 for (int i = 0; i < expectedSolution.length; ++i) {
75
76 final RealVector plusShift = MatrixUtils.createRealVector(expectedSolution);
77 plusShift.addToEntry(i, 2 * solutionTolerance);
78 boolean plusIsFeasible = true;
79 for (final Constraint constraint : constraints) {
80 plusIsFeasible &= constraint.overshoot(constraint.value(plusShift)) <= 0;
81 }
82 if (plusIsFeasible) {
83 // the plusShift point fulfills all constraints,
84 // so it must have worst objective function than the expected optimum
85 Assert.assertTrue(objectiveFunction.getObjectiveFunction().value(plusShift.toArray()) > expectedValue);
86 }
87
88 final RealVector minusShift = MatrixUtils.createRealVector(expectedSolution);
89 minusShift.addToEntry(i, -2 * solutionTolerance);
90 boolean minusIsFeasible = true;
91 for (final Constraint constraint : constraints) {
92 minusIsFeasible &= constraint.overshoot(constraint.value(minusShift)) <= 0;
93 }
94 if (minusIsFeasible) {
95 // the minusShift point fulfills all constraints,
96 // so it must have worst objective function than the expected optimum
97 Assert.assertTrue(objectiveFunction.getObjectiveFunction().value(minusShift.toArray()) > expectedValue);
98 }
99
100 }
101
102 }
103
104 }