ReferenceVectorAdaptation.h

Go to the documentation of this file.

//===========================================================================
/*!
 *
 *
 * \brief       Reference vector adaptation for the RVEA algorithm.
 *
 * \author      Bjoern Bugge Grathwohl
 * \date        March 2017
 *
 * \par Copyright 1995-2017 Shark Development Team
 *
 * <BR><HR>
 * This file is part of Shark.
 * <http://shark-ml.org/>
 *
 * Shark 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 3 of the License, or
 * (at your option) any later version.
 *
 * Shark 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 Shark.  If not, see <http://www.gnu.org/licenses/>.
 *
 */
//===========================================================================

#ifndef SHARK_ALGORITHMS_DIRECT_SEARCH_OPERATORS_REFERENCE_VECTOR_ADAPTATION_H
#define SHARK_ALGORITHMS_DIRECT_SEARCH_OPERATORS_REFERENCE_VECTOR_ADAPTATION_H

#include <shark/LinAlg/Base.h>
#include <shark/Algorithms/DirectSearch/Individual.h>

namespace shark {
/**
 * \brief Reference vector adaptation for the RVEA algorithm.
 *
 * This operator is supposed to be applied regularly in the RVEA algorithm to
 * adjust the set of reference vectors to better match a scaled pareto front.
 * See below paper for details:
 * R. Cheng, Y. Jin, M. Olhofer, and B. Sendhoff, “A reference vector guided
 * evolutionary algorithm for many-objective optimization,” IEEE Transactions on
 * Evolutionary Computation, Vol 20, Issue 5, October 2016
 * http://dx.doi.org/10.1109/TEVC.2016.2519378
 */
template <typename IndividualType>
struct ReferenceVectorAdaptation
{
    /**
     * \brief Apply adaptation operator and update the angles.
     */
    void operator()(
        std::vector<IndividualType> const & population,
        RealMatrix & referenceVectors,
        RealVector & minAngles)
    {
        auto f = unpenalizedFitness(population);
        const std::size_t sz = f.size();
        const std::size_t w = f[0].size();
        RealVector diff(w);
        for(std::size_t i = 0; i < w; ++i)
        {
            double max = std::numeric_limits<double>::min();
            double min = std::numeric_limits<double>::max();
            for(std::size_t j = 0; j < sz; ++j)
            {
                max = std::max(max, f[j][i]);
                min = std::min(min, f[j][i]);
            }
            double d = max - min;
            diff[i] = (d == 0) ? 1 : d;
        }
        referenceVectors = m_initVecs * repeat(diff, m_initVecs.size1());
        for(std::size_t i = 0; i < referenceVectors.size1(); ++i)
        {
            row(referenceVectors, i) /= norm_2(row(referenceVectors, i));
        }
        updateAngles(referenceVectors, minAngles);
    }

    /**
     * \brief Compute the minimum angles between unit vectors.
     */
    static void updateAngles(
        RealMatrix const & referenceVectors,
        RealVector & minAngles)
    {
        const std::size_t s = referenceVectors.size1();
        const RealMatrix m = acos(prod(referenceVectors,
                                       trans(referenceVectors))) +
            to_diagonal(RealVector(s, 1e10));
        for(std::size_t i = 0; i < s; ++i)
        {
            minAngles[i] = min(row(m, i));
        }
    }

    template <typename Archive>
    void serialize(Archive & archive)
    {
        archive & m_initVecs;
    }

    /// \brief The set of initial reference vectors.
    ///
    /// This must be set before the operator is called the first time.
    RealMatrix m_initVecs;
};

} // namespace shark

#endif // SHARK_ALGORITHMS_DIRECT_SEARCH_OPERATORS_REFERENCE_VECTOR_ADAPTATION_H