source: stable/HeuristicLab.Problems.Instances.DataAnalysis/3.3/Regression/Physics/AircraftMaximumLift.cs @ 17174

Last change on this file since 17174 was 17174, checked in by mkommend, 2 months ago

#3014: Merged r17092, r17093, r17094, r17150 into stable.

File size: 7.6 KB
Line 
1#region License Information
2/* HeuristicLab
3 * Copyright (C) 2002-2019 Heuristic and Evolutionary Algorithms Laboratory (HEAL)
4 *
5 * This file is part of HeuristicLab.
6 *
7 * HeuristicLab is free software: you can redistribute it and/or modify
8 * it under the terms of the GNU General Public License as published by
9 * the Free Software Foundation, either version 3 of the License, or
10 * (at your option) any later version.
11 *
12 * HeuristicLab is distributed in the hope that it will be useful,
13 * but WITHOUT ANY WARRANTY; without even the implied warranty of
14 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
15 * GNU General Public License for more details.
16 *
17 * You should have received a copy of the GNU General Public License
18 * along with HeuristicLab. If not, see <http://www.gnu.org/licenses/>.
19 */
20#endregion
21
22using System;
23using System.Collections.Generic;
24using System.Linq;
25using HeuristicLab.Common;
26using HeuristicLab.Random;
27
28namespace HeuristicLab.Problems.Instances.DataAnalysis {
29  public class AircraftMaximumLift : ArtificialRegressionDataDescriptor {
30    public override string Name { get { return "Aircraft Maximum Lift Coefficient f(X) = x1 - 1/4 x4 x5 x6 (4 + 0.1 x2/x3 - x2²/x3²) + x13 x14/x15 x18 x7 - x13 x14/x15 x8 + x13 x14/x15 x9 + x16 x17/x15 x18 x10 - x16 x17/x15 x11 + x16 x17/x15 x12"; } }
31
32    public override string Description {
33      get {
34        return "Slightly changed version of the problem instance given in: " + Environment.NewLine +
35          "Chen Chen, Changtong Luo, Zonglin Jiang, \"A multilevel block building algorithm for fast " +
36          "modeling generalized separable systems\", " +
37          "pre-print on arXiv: https://arxiv.org/abs/1706.02281 ." + Environment.NewLine +
38          "Notably, this problem is missing from the peer-reviewed version of the article in Expert Systems with Applications, Volume 109" + Environment.NewLine +
39          "Function: f(X) = x1 - 0.25 x4 x5 x6 (4 + 0.1 x2/x3 - x2²/x3²) + x13 x14/x15 x18 x7 - x13 x14/x15 x8 + x13 x14/x15 x9 + x16 x17/x15 x18 x10 - x16 x17/x15 x11 + x16 x17/x15 x12" + Environment.NewLine +
40          "with x1 ∈ [0.4, 0.8]," + Environment.NewLine +
41          "x2 ∈ [3, 4]," + Environment.NewLine +
42          "x3 ∈ [20, 30]," + Environment.NewLine +
43          "x4, x13, x16 ∈ [2, 5]," + Environment.NewLine +
44          "x14, x17 ∈ [1, 1.5]," + Environment.NewLine +
45          "x15 ∈ [5, 7]," + Environment.NewLine +
46          "x18 ∈ [10, 20]," + Environment.NewLine +
47          "x8, x11 ∈ [1, 1.5]," + Environment.NewLine +
48          "x9, x12 ∈ [1, 2]," + Environment.NewLine +
49          "x7, x10 ∈ [0.5, 1.5]." + Environment.NewLine +
50          "Values for x5 and x6 have not been specified in the reference paper." +
51          " We therefore only use a single (x5) variable in place of ∆αW/c and set x6 to a constant value of 1.0." + Environment.NewLine +
52          "The range for x5 is [0..20].";
53      }
54    }
55
56    protected override string TargetVariable { get { return "f(X)"; } }
57    protected override string[] VariableNames { get { return new string[] { "x1", "x2", "x3", "x4", "x5", "x6", "x7", "x8", "x9", "x10", "x11", "x12", "x13", "x14", "x15", "x16", "x17", "x18", "f(X)", "f(X)_noise" }; } }
58    protected override string[] AllowedInputVariables { get { return VariableNames.Except(new string[] { "f(X)", "f(X)_noise" }).ToArray(); } }
59    protected override int TrainingPartitionStart { get { return 0; } }
60    protected override int TrainingPartitionEnd { get { return 100; } }
61    protected override int TestPartitionStart { get { return 100; } }
62    protected override int TestPartitionEnd { get { return 200; } }
63
64    public int Seed { get; private set; }
65
66    public AircraftMaximumLift() : this((int)System.DateTime.Now.Ticks) { }
67
68    public AircraftMaximumLift(int seed) {
69      Seed = seed;
70    }
71
72    protected override List<List<double>> GenerateValues() {
73      var rand = new MersenneTwister((uint)Seed);
74
75      List<List<double>> data = new List<List<double>>();
76      var x1 = ValueGenerator.GenerateUniformDistributedValues(rand.Next(), TestPartitionEnd, 0.4, 0.8).ToList();
77
78      var x2 = ValueGenerator.GenerateUniformDistributedValues(rand.Next(), TestPartitionEnd, 3.0, 4.0).ToList();
79
80      var x3 = ValueGenerator.GenerateUniformDistributedValues(rand.Next(), TestPartitionEnd, 20.0, 30.0).ToList();
81
82      var x4 = ValueGenerator.GenerateUniformDistributedValues(rand.Next(), TestPartitionEnd, 2.0, 5.0).ToList();
83      var x13 = ValueGenerator.GenerateUniformDistributedValues(rand.Next(), TestPartitionEnd, 2.0, 5.0).ToList();
84      var x16 = ValueGenerator.GenerateUniformDistributedValues(rand.Next(), TestPartitionEnd, 2.0, 5.0).ToList();
85
86
87      // in the reference paper \Delta alpha_w/c is replaced by two variables x5*x6.
88      var x5 = ValueGenerator.GenerateUniformDistributedValues(rand.Next(), TestPartitionEnd, 0, 20).ToList(); // range for X5 is not specified in the paper, we use [0°..20°] for ∆αW/c
89      var x6 = Enumerable.Repeat(1.0, x5.Count).ToList(); // range for X6 is not specified in the paper. In the maximum lift formular there is only a single variable ∆αW/c in place of x5*x6.
90
91      var x7 = ValueGenerator.GenerateUniformDistributedValues(rand.Next(), TestPartitionEnd, 0.5, 1.5).ToList();
92      var x10 = ValueGenerator.GenerateUniformDistributedValues(rand.Next(), TestPartitionEnd, 0.5, 1.5).ToList();
93
94      var x8 = ValueGenerator.GenerateUniformDistributedValues(rand.Next(), TestPartitionEnd, 1.0, 1.5).ToList();
95      var x11 = ValueGenerator.GenerateUniformDistributedValues(rand.Next(), TestPartitionEnd, 1.0, 1.5).ToList();
96
97      var x9 = ValueGenerator.GenerateUniformDistributedValues(rand.Next(), TestPartitionEnd, 1.0, 2.0).ToList();
98      var x12 = ValueGenerator.GenerateUniformDistributedValues(rand.Next(), TestPartitionEnd, 1.0, 2.0).ToList();
99
100      var x14 = ValueGenerator.GenerateUniformDistributedValues(rand.Next(), TestPartitionEnd, 1.0, 1.5).ToList();
101      var x17 = ValueGenerator.GenerateUniformDistributedValues(rand.Next(), TestPartitionEnd, 1.0, 1.5).ToList();
102
103      var x15 = ValueGenerator.GenerateUniformDistributedValues(rand.Next(), TestPartitionEnd, 5.0, 7.0).ToList();
104
105      var x18 = ValueGenerator.GenerateUniformDistributedValues(rand.Next(), TestPartitionEnd, 10.0, 20.0).ToList();
106
107
108      List<double> fx = new List<double>();
109      List<double> fx_noise = new List<double>();
110      data.Add(x1);
111      data.Add(x2);
112      data.Add(x3);
113      data.Add(x4);
114      data.Add(x5);
115      data.Add(x6);
116      data.Add(x7);
117      data.Add(x8);
118      data.Add(x9);
119      data.Add(x10);
120      data.Add(x11);
121      data.Add(x12);
122      data.Add(x13);
123      data.Add(x14);
124      data.Add(x15);
125      data.Add(x16);
126      data.Add(x17);
127      data.Add(x18);
128      data.Add(fx);
129      data.Add(fx_noise);
130
131      for (int i = 0; i < x1.Count; i++) {
132        double fxi = x1[i];
133        fxi = fxi - 0.25 * x4[i] * x5[i] * x6[i] * (4 + 0.1 * (x2[i] / x3[i]) - (x2[i] / x3[i]) * (x2[i] / x3[i]));
134        fxi = fxi + x13[i] * (x14[i] / x15[i]) * x18[i] * x7[i];
135        fxi = fxi - x13[i] * (x14[i] / x15[i]) * x8[i];
136        fxi = fxi + x13[i] * (x14[i] / x15[i]) * x9[i];
137        fxi = fxi + x16[i] * (x17[i] / x15[i]) * x18[i] * x10[i];
138        fxi = fxi - x16[i] * (x17[i] / x15[i]) * x11[i];
139        fxi = fxi + x16[i] * (x17[i] / x15[i]) * x12[i];
140
141        fx.Add(fxi);
142      }
143
144      var sigma_noise = 0.05 * fx.StandardDeviationPop();
145      fx_noise.AddRange(fx.Select(fxi => fxi + NormalDistributedRandom.NextDouble(rand, 0, sigma_noise)));
146
147      return data;
148    }
149  }
150}
Note: See TracBrowser for help on using the repository browser.