evaluationop.hpp

非常好的进化算法EC 实现平台 可以实现多种算法 GA GP

/*
 *  Open BEAGLE
 *  Copyright (C) 2001-2005 by Christian Gagne and Marc Parizeau
 *
 *  This library is free software; you can redistribute it and/or
 *  modify it under the terms of the GNU Lesser General Public
 *  License as published by the Free Software Foundation; either
 *  version 2.1 of the License, or (at your option) any later version.
 *
 *  This library 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
 *  Lesser General Public License for more details.
 *
 *  You should have received a copy of the GNU Lesser General Public
 *  License along with this library; if not, write to the Free Software
 *  Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
 *
 *  Contact:
 *  Laboratoire de Vision et Systemes Numeriques
 *  Departement de genie electrique et de genie informatique
 *  Universite Laval, Quebec, Canada, G1K 7P4
 *  http://vision.gel.ulaval.ca
 *
 */

/*!
 *  \file   beagle/Coev/EvaluationOp.hpp
 *  \brief  Declaration of file Beagle::Coev::EvaluationOp.
 *  \author Christian Gagne
 *  \author Marc Parizeau
 *  $Revision: 1.17 $
 *  $Date: 2005/10/04 09:32:52 $
 */

#ifndef Beagle_Coev_EvaluationOp_hpp
#define Beagle_Coev_EvaluationOp_hpp

#include <string>

#include "Threading.hpp"

#include "beagle/config.hpp"
#include "beagle/macros.hpp"
#include "beagle/Object.hpp"
#include "beagle/AbstractAllocT.hpp"
#include "beagle/PointerT.hpp"
#include "beagle/ContainerT.hpp"
#include "beagle/EvaluationOp.hpp"


namespace Beagle {
namespace Coev {

/*!
 *  \class EvaluationOp beagle/Coev/EvaluationOp.hpp "beagle/Coev/EvaluationOp.hpp"
 *  \brief Abstract co-evolution evaluation operator class.
 *  \ingroup COEVF
 *
 *  For its application co-evolution fitness evaluation operators, the user must over-defined the pure 
 *  virtual methods makeSets and evaluateSets. The makeSets method is called to ask the evaluation
 *  operator to make evaluation sets from the actual deme. An evaluation set corresponds to a 
 *  group of individuals than can be mated with groups of individual from other population for
 *  co-evolutionary fitness evaluation. The groups of individuals created in method makeSets are 
 *  added to the shared structure by a call to the method addSets. This method takes a second
 *  optionnal argument which indicated whether or not the call is blocking. If the call is 
 *  blocking, the call to addSets will return only when the set will be evaluated. Otherwize, the
 *  evaluation set return immediatly after the set is added  to the shared structure. Once the 
 *  number of evaluation sets added to the shared structure is equal to the trigger value, the
 *  problem-specific co-evolutionary fitness evaluation is done by a call to the pure virtual method 
 *  evaluateSets.
 *
 *  Fitness value evaluated in method evaluateSets must be assigned to the individuals using
 *  the method assignFitness. If this method is not used, the statistics values will
 *  not be accurate.
 *
 *  The member mID of the struct EvalSet is provided for convenience only and is not used internally.
 *  This can be used to identity the different evaluation sets add into the common structure. You 
 *  cannot make any assumptions on the order the evaluation sets are put in the common structure, 
 *  and so the order they are given to method evaluateSets.
 *
 *  The trigger value is specified when constructing the different co-evolutionary evaluation operator.
 *  Each co-evolutionary evaluation must specify the trigger value, and the value given must be the 
 *  same for the different thread/population evaluation operators.
 *
 */
class EvaluationOp : public Beagle::EvaluationOp {

public:

  //! Coev::EvaluationOp allocator type.
  typedef AbstractAllocT<EvaluationOp,Beagle::EvaluationOp::Alloc>
          Alloc;
  //! Coev::EvaluationOp handle type.
  typedef PointerT<EvaluationOp,Beagle::EvaluationOp::Handle>
          Handle;
  //! Coev::EvaluationOp bag type.
  typedef ContainerT<EvaluationOp,Beagle::EvaluationOp::Bag>
          Bag;

  /*!
   *  \struct EvalSet beagle/Coev/EvaluationOp.hpp "beagle/Coev/EvaluationOp.hpp"
   *  \brief Evaluation set association used in co-evolution evaluation operator.
   */
  struct EvalSet {
    Beagle::Individual::Bag mIndividuals;    //!< Individuals of the evaluation set
    Beagle::Context::Handle mContext;        //!< Context associated to the evaluation set
    unsigned int            mID;             //!< Evaluation set ID

    EvalSet();
    EvalSet(Beagle::Individual::Bag& inIndividuals,
            Beagle::Context::Handle inContext,
            unsigned int inID=0);
  };

  //! Vector of evaluation set.
  typedef std::vector< EvalSet,BEAGLE_STLALLOCATOR<EvalSet> >
          EvalSetVector;

  explicit EvaluationOp(unsigned int inTrigger=1, string inName="Coev-EvaluationOp");
  virtual ~EvaluationOp() { }

  /*!
   *  \brief Evaluate fitness of a bunch of individual in an evaluation sets.
   *  \param ioSets Sets to evaluate fitness.
   *
   *  You cannot make assumption about the order of the evaluation sets are in
   *  the ioSets structure. These can change from evaluation to evaluation. Use
   *  struct EvalSet member inID to identify the evaluation sets you put into
   *  shared storage by call to add set. This ID is for convenience and is not used
   *  internally to identify sets.
   *
   *  Assign fitness value to individuals using method assignFitness, otherwise
   *  statistics value will be erroneous.
   */
  virtual void evaluateSets(EvalSetVector& ioSets) =0;

  /*!
   *  \brief Make evaluation sets from given deme.
   *  \param ioIndivBag Bag of individuals to evaluate.
   *  \param ioContext Evolutionary context.
   *
   *  This method is call one time at each generation, for each evolving
   *  thread/population. This method is problem-specific and consists
   *  to make evaluation sets that are mated with other thread/population
   *  evaluation sets for co-evolutionary fitness evaluation. Sets are added
   *  into a shared structure with a call to method addSet.
   */
  virtual void makeSets(Beagle::Individual::Bag& ioIndivBag,
                        Beagle::Context::Handle ioContext) =0;

  virtual void addSet(EvalSet& inEvalSet, bool inBlocking=true);
  virtual void assignFitness(Beagle::Fitness::Handle inFitness,
                             Beagle::Individual& ioIndividual,
                             Beagle::Context& ioContext) const;
  virtual Individual::Handle breed(Individual::Bag& inBreedingPool,
                                   BreederNode::Handle inChild,
                                   Context& ioContext);
  virtual void operate(Beagle::Deme& ioDeme, Beagle::Context& ioContext);

protected:

  static PACC::Threading::Condition smCondition;      //!< Condition of co-evaluation
  static EvalSetVector              smEvalSets;       //!< Shared storage of evaluation sets
  static unsigned int               smTrigger;        //!< Number of sets needed to start an evaluation

private:

  virtual Fitness::Handle evaluate(Individual& inIndividual, Context& ioContext);

};

}
}


#endif // Beagle_Coev_EvaluationOp_hpp