eu.crisis_economics.abm.markets.clearing.heterogeneous.LevenbergMarquardtClearingAlgorithm.java Source code

Introduction

Here is the source code for eu.crisis_economics.abm.markets.clearing.heterogeneous.LevenbergMarquardtClearingAlgorithm.java

Source

/*
 * This file is part of CRISIS, an economics simulator.
 * 
 * Copyright (C) 2015 John Kieran Phillips
 *
 * CRISIS 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.
 *
 * CRISIS 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 CRISIS.  If not, see <http://www.gnu.org/licenses/>.
 */
package eu.crisis_economics.abm.markets.clearing.heterogeneous;

import org.apache.commons.math3.fitting.leastsquares.LeastSquaresFactory;
import org.apache.commons.math3.fitting.leastsquares.LeastSquaresOptimizer.Optimum;
import org.apache.commons.math3.fitting.leastsquares.LeastSquaresProblem;
import org.apache.commons.math3.fitting.leastsquares.LevenbergMarquardtOptimizer;
import org.apache.commons.math3.fitting.leastsquares.MultivariateJacobianFunction;
import org.apache.commons.math3.linear.ArrayRealVector;
import org.apache.commons.math3.linear.RealVector;
import org.apache.commons.math3.optim.ConvergenceChecker;
import org.apache.commons.math3.optim.SimpleVectorValueChecker;

import com.google.common.base.Preconditions;

/**
  * An implementation of the Levenberg-Marquardt SSQ fitting algorithm
  * for mixed network clearing.
  * 
  * @author phillips
  */
final class LevenbergMarquardtClearingAlgorithm extends NumericalDerivativeClearingAlgorithm {

    private final int maximumIterations, maximumEvaluations;
    private final double absErrorTarget, relErrorTarget;

    public LevenbergMarquardtClearingAlgorithm( // Immutable
            final int maximumIterations, final int maximumEvaluations, final double absErrorTarget,
            final double relErrorTarget) {
        this.maximumIterations = maximumIterations;
        this.maximumEvaluations = maximumEvaluations;
        this.absErrorTarget = absErrorTarget;
        this.relErrorTarget = relErrorTarget;
    }

    @Override
    public double applyToNetwork(final MixedClearingNetwork network) {
        Preconditions.checkNotNull(network);

        final VectorCostFunction function = super.getVectorCostFunction(network);
        final MultivariateJacobianFunction model = LeastSquaresFactory.model(function,
                super.getJacobianMatrixFunction(network));
        final RealVector observed = new ArrayRealVector(super.calculateTarget(network)),
                start = new ArrayRealVector(network.getNumberOfEdges());
        for (int i = 0; i < network.getNumberOfEdges(); ++i)
            start.setEntry(i, network.getEdges().get(i).getMaximumRateAdmissibleByBothParties());
        start.set(1.0);

        final ConvergenceChecker<LeastSquaresProblem.Evaluation> evaluationChecker = LeastSquaresFactory
                .evaluationChecker(new SimpleVectorValueChecker(relErrorTarget, absErrorTarget));
        final LeastSquaresProblem problem = LeastSquaresFactory.create(model, observed, start, evaluationChecker,
                maximumEvaluations, maximumIterations);

        final LevenbergMarquardtOptimizer optimizer = new LevenbergMarquardtOptimizer();
        final Optimum result = optimizer.optimize(problem);

        final double residualCost = result.getRMS();
        System.out.println("Network cleared: residual cost: " + residualCost + ".");

        return residualCost;
    }

    /**
      * @return The maximum number of Levenberg iterations.
      */
    public int getMaximumIterations() {
        return maximumIterations;
    }

    /**
     * @return The maximum number of network cost evaluations.
     */
    public int getMaximumEvaluations() {
        return maximumEvaluations;
    }

    /**
     * @return The absolute residual error threshold target.
     */
    public double getAbsErrorTarget() {
        return absErrorTarget;
    }

    /**
     * @return The relative residual error threshold target.
     */
    public double getRelErrorTarget() {
        return relErrorTarget;
    }

    /*
     * Returns a brief description of this object. The exact details of the
     * string are subject to change, and as such should not be regarded as fixed.
     */
    @Override
    public String toString() {
        return "Levenberg Marquardt Clearing Algorithm, maximum iterations:" + maximumIterations
                + ", maximum network cost evaluations:" + maximumEvaluations + ", absolute error target:"
                + absErrorTarget + ", relative error target:" + relErrorTarget + ".";
    }
}