source: branches/DataAnalysisService/HeuristicLab.Encodings.ParameterConfigurationTreeEncoding/3.3/MetaOptimizationUtil.cs @ 9716

#region License Information
/* HeuristicLab
 * Copyright (C) 2002-2012 Heuristic and Evolutionary Algorithms Laboratory (HEAL)
 *
 * This file is part of HeuristicLab.
 *
 * HeuristicLab is free software: you can redistribute it and/or modify
 * it under the terms of the GNU General Public License as published by
 * the Free Software Foundation, either version 3 of the License, or
 * (at your option) any later version.
 *
 * HeuristicLab is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 * GNU General Public License for more details.
 *
 * You should have received a copy of the GNU General Public License
 * along with HeuristicLab. If not, see <http://www.gnu.org/licenses/>.
 */
#endregion

using System;
using System.Collections.Generic;
using System.Linq;
using HeuristicLab.Core;
using HeuristicLab.Data;
using HeuristicLab.Optimization;

namespace HeuristicLab.Encodings.ParameterConfigurationTreeEncoding {
  public static class MetaOptimizationUtil {
    /// <summary>
    /// Removes those results from the run which are not declared in resultsToKeep
    /// </summary>
    public static void ClearResults(IRun run, IEnumerable<string> resultsToKeep) {
      var resultsToRemove = new List<string>();
      foreach (var result in run.Results) {
        if (!resultsToKeep.Contains(result.Key))
          resultsToRemove.Add(result.Key);
      }
      foreach (var result in resultsToRemove)
        run.Results.Remove(result);
    }

    /// <summary>
    /// Removes those parameters from the run which are not declared in parametersToKeep
    /// </summary>
    public static void ClearParameters(IRun run, IEnumerable<string> parametersToKeep) {
      var parametersToRemove = new List<string>();
      foreach (var parameter in run.Parameters) {
        if (!parametersToKeep.Contains(parameter.Key))
          parametersToRemove.Add(parameter.Key);
      }
      foreach (var parameter in parametersToRemove)
        run.Parameters.Remove(parameter);
    }

    public static double Normalize(
          ParameterConfigurationTree parameterConfigurationTree,
          double[] referenceQualityAverages,
          double[] referenceQualityDeviations,
          double[] referenceEvaluatedSolutionAverages,
          double qualityAveragesWeight,
          double qualityDeviationsWeight,
          double evaluatedSolutionsWeight,
          bool maximization) {

      double[] qualityAveragesNormalized = new double[referenceQualityAverages.Length];
      double[] qualityDeviationsNormalized = new double[referenceQualityDeviations.Length];
      double[] evaluatedSolutionAveragesNormalized = new double[referenceEvaluatedSolutionAverages.Length];

      for (int i = 0; i < referenceQualityAverages.Length; i++) {
        qualityAveragesNormalized[i] = parameterConfigurationTree.AverageQualities[i] / referenceQualityAverages[i];
        qualityDeviationsNormalized[i] = parameterConfigurationTree.QualityStandardDeviations[i] / referenceQualityDeviations[i];
        evaluatedSolutionAveragesNormalized[i] = parameterConfigurationTree.AverageEvaluatedSolutions[i] / referenceEvaluatedSolutionAverages[i];
      }
      parameterConfigurationTree.NormalizedQualityAverages = new DoubleArray(qualityAveragesNormalized);
      parameterConfigurationTree.NormalizedQualityDeviations = new DoubleArray(qualityDeviationsNormalized);
      parameterConfigurationTree.NormalizedEvaluatedSolutions = new DoubleArray(evaluatedSolutionAveragesNormalized);

      double qualityAveragesNormalizedValue = qualityAveragesNormalized.Average();
      double qualityDeviationsNormalizedValue = qualityDeviationsNormalized.Average();
      double evaluatedSolutionAveragesNormalizedValue = evaluatedSolutionAveragesNormalized.Average();

      // deviation and evaluatedSolutions are always minimization problems. so if maximization=true, flip the values around 1.0 (e.g. 1.15 -> 0.85)
      if (maximization) {
        qualityDeviationsNormalizedValue -= (qualityDeviationsNormalizedValue - 1) * 2;
        evaluatedSolutionAveragesNormalizedValue -= (evaluatedSolutionAveragesNormalizedValue - 1) * 2;
      }

      // apply weights
      qualityAveragesNormalizedValue *= qualityAveragesWeight;
      qualityDeviationsNormalizedValue *= qualityDeviationsWeight;
      evaluatedSolutionAveragesNormalizedValue *= evaluatedSolutionsWeight;

      double weightSum = qualityAveragesWeight + qualityDeviationsWeight + evaluatedSolutionsWeight;
      parameterConfigurationTree.Quality = new DoubleValue((qualityAveragesNormalizedValue + qualityDeviationsNormalizedValue + evaluatedSolutionAveragesNormalizedValue) / weightSum);

      return parameterConfigurationTree.Quality.Value;
    }

    /// <summary>
    /// Creates a new instance of algorithmType, sets the given problem and parameterizes it with the given configuration
    /// </summary>
    public static IAlgorithm CreateParameterizedAlgorithmInstance(ParameterConfigurationTree parameterConfigurationTree, Type algorithmType, IProblem problem, bool randomize = false, IRandom random = null) {
      var algorithm = (IAlgorithm)Activator.CreateInstance(algorithmType);
      algorithm.Problem = problem;
      if (algorithm is EngineAlgorithm) {
        ((EngineAlgorithm)algorithm).Engine = new SequentialEngine.SequentialEngine();
      }
      if (randomize) parameterConfigurationTree.Randomize(random);
      parameterConfigurationTree.Parameterize(algorithm);
      return algorithm;
    }
  }
}