GarpAlgorithm.h

Go to the documentation of this file.

/* **************************************
 *  GARP Modeling Package
 *
 * **************************************
 *
 * Copyright (c), The Center for Research, University of Kansas, 2385 Irving Hill Road, Lawrence, KS 66044-4755, USA.
 * Copyright (C), David R.B. Stockwell of Symbiotik Pty. Ltd.
 *
 * This program is free software; you can redistribute it and/or modify
 * it under the terms of the license that is distributed with the software.
 *
 * This program 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
 * license.txt file included with this software for more details.
 */

// GarpAlgorithm.h : Declaration of the Algorithm class

#ifndef __GARPALGORITHM_H_
#define __GARPALGORITHM_H_

#include "RuleSet.h"
#include "Utilities.h"

#include <openmodeller/om.hh>

class EnvCellSet;

// ========================================================================
// GarpAlgorithm
// ========================================================================
class GarpAlgorithm : public AlgorithmImpl
{
private:
  // current ruleset
  RuleSet objNew;

  // best model so far
  RuleSet objBest;

  // data points used for training
  EnvCellSet * objTrainSet;


public:
  // OM Algorithm virtual methods
  int initialize();
  int iterate();
  int done() const;
  float getProgress() const;
  Scalar getValue( const Sample& x ) const;
  int getConvergence( Scalar * const val ) const;
  int getGeneration() { return Gen; }


  GarpAlgorithm();
  virtual ~GarpAlgorithm();

  RuleSet * getBestSet() { return &objBest; }
  void getInitialModel(int intSize, EnvCellSet * objTrainSet);
  void updateRuleUsage();

  int ruleSetSize()
  { return objBest.size(); }

  char * getParameter2(char * sParamName);
  void setParameter(char * sParamName, char * sParamValue);

  char * getSelectedLayersAsString();
  void setSelectedLayers(char * strParamValue);

 protected:
  void _getConfiguration( ConfigurationPtr& ) const;
  void _setConfiguration( const ConstConfigurationPtr& );

private:
  // private methods
  void initializeProperties();

  // Garp algorithm
  void DisplayStatus();
  void colonize(RuleSet * objRules, EnvCellSet * objTrainSet, int intNewRules);
  void evaluate(RuleSet * objRules, EnvCellSet * objTrainSet);
  int saveRule(int iIndex);
  void measure();
  double converge();
  void select();

  void mutate();
  void crossover();
  void join();

  void updateHeuOpPerformance(char chrType);

  void generate(EnvCellSet * objTestDataset);

  // deprecated methods
  void saveBestRules(RuleSet * toRuleSet, RuleSet * fromRuleSet);
  void concatenateRuleSets(RuleSet * toRuleSet, RuleSet * fromRuleSet);


private:

  long lVersion;

  // flags
  int Sigflag;
  int Postflag;
  int Compflag;
  int Adjustflag;

  int BioclimOnlyFlag;
  int LogitOnlyFlag;  
  int RangeRuleFlag;  
  int NegatedRuleFlag;
  int AtomicRuleFlag; 
  int LogitRuleFlag;  

  bool bGeneIsActive[MAX_ENV_LAYERS];
  int iGeneIndex[MAX_ENV_LAYERS];
  int iActiveGenes;

  // timming
  int iCPUTime;         // accumulated CPU time used for processing the model (in miliseconds) 

  // algorithm variables
  int Resamples;

  double Accuracylimit; // The minimium post prob of a rule
  double MinUsage;
  double Mortality;
  
  int Totalgens;    // generations per experiment
  int Totaltrials;      // trials per experiment               
  int Popsize;        // population size
  double C_rate;       // crossover rate                      
  double M_rate;       // mutation rate                       
  double J_rate;       // join rate
  double I_rate;
  double Gapsize;   // fraction of pop generated from archive
  int Trials;          // trial counter
  int Maxspin;          // max gens without evals
  double Resampling_f;
  double Significance;   // the minimum level for inc in best
  double Conv_limit;    // the fractional addition of rules
  double Cutval;


  // data collection and loop control variables
  double  Ave_current_perf;/* ave perf in current generation       */
  double  Best;            /* best performance seen so far         */
  double  Best_current_perf;/* best perf in current generation     */
  int    Best_guy;        /* index of best_current_perf           */
  int    Conv;            /* number of partially coverged genes   */
  mutable bool   Doneflag;        /* set when termination conditions hold */
  int    Experiment;      /* experiment counter                   */
  int    Gen;             /* generation counter                   */
  int    Lost;            /* number of totally coverged positions */
  int    Spin;            /* number of gens since eval occurred   */
  double  Worst;           /* worst performance seen so far        */
  double  Worst_current_perf;/* worst perf in current generation   */
  int     Heuristic[2][5];        /* successes of heuristic operators */
  double  Convergence;        /* the stability of rule set */
  int     Improvements;   /* the number of times the best set has been 
                improved by addition or alteration */
  int Resample;

  mutable float _maxProgress;

};

// ========================================================================

#endif //__GarpAlgorithm_H_