/*
 * <moeoExpBinaryIndicatorBasedFitnessAssignment.h>
 * Copyright (C) DOLPHIN Project-Team, INRIA Futurs, 2006-2014
 * (C) OPAC Team, LIFL, 2002-2014
 *
 * Arnaud Liefooghe
 *
 * Oct 17, 2014  -   Arnaud Liefooghe
 *   Modifications on the handling of the internal data structure (values)
 *   in order to avoid (too much/bad) memory consumptions, in particular
 *   when a very large population size is used
 *
 * This software is governed by the CeCILL license under French law and
 * abiding by the rules of distribution of free software.  You can  use,
 * modify and/ or redistribute the software under the terms of the CeCILL
 * license as circulated by CEA, CNRS and INRIA at the following URL
 * "http://www.cecill.info".
 *
 * As a counterpart to the access to the source code and  rights to copy,
 * modify and redistribute granted by the license, users are provided only
 * with a limited warranty  and the software's author,  the holder of the
 * economic rights,  and the successive licensors  have only  limited liability.
 *
 * In this respect, the user's attention is drawn to the risks associated
 * with loading,  using,  modifying and/or developing or reproducing the
 * software by the user in light of its specific status of free software,
 * that may mean  that it is complicated to manipulate,  and  that  also
 * therefore means  that it is reserved for developers  and  experienced
 * professionals having in-depth computer knowledge. Users are therefore
 * encouraged to load and test the software's suitability as regards their
 * requirements in conditions enabling the security of their systems and/or
 * data to be ensured and,  more generally, to use and operate it in the
 * same conditions as regards security.
 * The fact that you are presently reading this means that you have had
 * knowledge of the CeCILL license and that you accept its terms.
 *
 * ParadisEO WebSite : http://paradiseo.gforge.inria.fr
 * Contact: paradiseo-help@lists.gforge.inria.fr
 *
 */
//-----------------------------------------------------------------------------

#ifndef MOEOEXPBINARYINDICATORBASEDFITNESSASSIGNMENT_H_
#define MOEOEXPBINARYINDICATORBASEDFITNESSASSIGNMENT_H_

#include <math.h>
#include <vector>
#include <eoPop.h>
#include <fitness/moeoBinaryIndicatorBasedFitnessAssignment.h>
#include <metric/moeoNormalizedSolutionVsSolutionBinaryMetric.h>
#include <utils/moeoConvertPopToObjectiveVectors.h>

/**
 * Fitness assignment sheme based on an indicator proposed in:
 * E. Zitzler, S. Künzli, "Indicator-Based Selection in Multiobjective Search", Proc. 8th International Conference on Parallel Problem Solving from Nature (PPSN VIII), pp. 832-842, Birmingham, UK (2004).
 * This strategy is, for instance, used in IBEA.
 */
template < class MOEOT >
class moeoExpBinaryIndicatorBasedFitnessAssignment : public moeoBinaryIndicatorBasedFitnessAssignment < MOEOT >
{
public:
    
    /** The type of objective vector */
    typedef typename MOEOT::ObjectiveVector ObjectiveVector;
    typedef typename ObjectiveVector::Type Type;
    typedef typename MOEOT::Fitness Fitness;
    
    
    /**
     * Ctor.
     * @param _metric the quality indicator
     * @param _kappa the scaling factor
     */
    moeoExpBinaryIndicatorBasedFitnessAssignment(moeoNormalizedSolutionVsSolutionBinaryMetric < ObjectiveVector, double > & _metric, const double _kappa = 0.05) : metric(_metric), kappa(_kappa), values(0) {}
    
    
    /**
     * Dtor.
     */
    ~moeoExpBinaryIndicatorBasedFitnessAssignment()
    {
        // clear "values"
        for (unsigned int i=0; i<values.size(); i++) values[i].clear();
        values.clear();
    }
    
    
    /**
     * Sets the fitness values for every solution contained in the population _pop
     * @param _pop the population
     */
    virtual void operator()(eoPop < MOEOT > & _pop)
    {
        // 1 - setting of the bounds
        setup(_pop);
        // 2 - computing every indicator values
        computeValues(_pop);
        // 3 - setting fitnesses
        setFitnesses(_pop);
    }
    
    
    /**
     * Updates the fitness values of the whole population _pop by taking the deletion of the objective vector _objVec into account.
     * @param _pop the population
     * @param _objVec the objective vector
     */
    void updateByDeleting(eoPop < MOEOT > & _pop, ObjectiveVector & _objVec)
    {
        std::vector < double > v;
        v.resize(_pop.size());
        for (unsigned int i=0; i<_pop.size(); i++)
        {
            v[i] = metric(_objVec, _pop[i].objectiveVector());
        }
        for (unsigned int i=0; i<_pop.size(); i++)
        {
            _pop[i].fitness( _pop[i].fitness() + exp(-v[i]/kappa) );
        }
        v.clear();
    }
    
    
    /**
     * Updates the fitness values of the whole population _pop by taking the adding of the objective vector _objVec into account
     * and returns the fitness value of _objVec.
     * @param _pop the population
     * @param _objVec the objective vector
     */
    double updateByAdding(eoPop < MOEOT > & _pop, ObjectiveVector & _objVec)
    {
        std::vector < double > v;
        // update every fitness values to take the new individual into account
        v.resize(_pop.size());
        for (unsigned int i=0; i<_pop.size(); i++)
        {
            v[i] = metric(_objVec, _pop[i].objectiveVector());
        }
        for (unsigned int i=0; i<_pop.size(); i++)
        {
            _pop[i].fitness( _pop[i].fitness() - exp(-v[i]/kappa) );
        }
        // compute the fitness of the new individual
        for (unsigned int i=0; i<_pop.size(); i++)
        {
            v[i] = metric(_pop[i].objectiveVector(), _objVec);
        }
        double result = 0;
        for (unsigned int i=0; i<v.size(); i++)
        {
            result -= exp(-v[i]/kappa);
        }
        v.clear();
        return result;
    }
    
    
protected:
    
    /** the quality indicator */
    moeoNormalizedSolutionVsSolutionBinaryMetric < ObjectiveVector, double > & metric;
    /** the scaling factor */
    double kappa;
    /** the computed indicator values */
    std::vector < std::vector<Type> > values;
    
    
    /**
     * Sets the bounds for every objective using the min and the max value for every objective vector of _pop
     * @param _pop the population
     */
    void setup(const eoPop < MOEOT > & _pop)
    {
        typename MOEOT::ObjectiveVector::Type min, max;
        for (unsigned int i=0; i<ObjectiveVector::Traits::nObjectives(); i++)
        {
            min = _pop[0].objectiveVector()[i];
            max = _pop[0].objectiveVector()[i];
            for (unsigned int j=1; j<_pop.size(); j++)
            {
                min = std::min(min, _pop[j].objectiveVector()[i]);
                max = std::max(max, _pop[j].objectiveVector()[i]);
            }
            // setting of the bounds for the objective i
            metric.setup(min, max, i);
        }
    }
    
    
    /**
     * Compute every indicator value in values (values[i] = I(_v[i], _o))
     * @param _pop the population
     */
    virtual void computeValues(const eoPop < MOEOT > & _pop)
    {
        // initialize the values structure (if it is used for the first time with such a population size)
        if (values.size() < _pop.size())
        {
            values.resize(_pop.size());
            for (unsigned int i=0; i<values.size(); i++)
            {
                values[i].resize(_pop.size());
            }
        }
        // go
        for (unsigned int i=0; i<_pop.size(); i++)
        {
            // the metric may not be symetric, thus neither is the matrix
            for (unsigned int j=0; j<_pop.size(); j++)
            {
                if (i != j)
                {
                    values[i][j] = Type( metric(_pop[i].objectiveVector(), _pop[j].objectiveVector()) );
                }
            }
        }
    }
    
    
    /**
     * Sets the fitness value of the whole population
     * @param _pop the population
     */
    virtual void setFitnesses(eoPop < MOEOT > & _pop)
    {
        for (unsigned int i=0; i<_pop.size(); i++)
        {
            _pop[i].fitness(computeFitness(_pop, i));
        }
    }
    
    
    /**
     * Returns the fitness value of the _idx th individual of the population
     * @param _pop the population (only useful for its size here)
     * @param _idx the index
     */
    virtual Fitness computeFitness(const eoPop < MOEOT > & _pop, const unsigned int _idx)
    {
        Fitness result(0.0);
        for (unsigned int i=0; i<_pop.size(); i++)
        {
            if (i != _idx)
            {
                result -= exp(-values[i][_idx]/kappa);
            }
        }
        return result;
    }
    
};

#endif /*MOEOEXPBINARYINDICATORBASEDFITNESSASSIGNMENT_H_*/