org.cicirello.search.evo.BoltzmannStochasticUniversalSampling Maven / Gradle / Ivy

Go to download

Show more of this group Show more artifacts with this name
Show all versions of chips-n-salsa Show documentation

Chips-n-Salsa is a Java library of customizable, hybridizable, iterative, parallel, stochastic, and self-adaptive local search algorithms. The library includes implementations of several stochastic local search algorithms, including simulated annealing, hill climbers, as well as constructive search algorithms such as stochastic sampling. Chips-n-Salsa now also includes genetic algorithms as well as evolutionary algorithms more generally. The library very extensively supports simulated annealing. It includes several classes for representing solutions to a variety of optimization problems. For example, the library includes a BitVector class that implements vectors of bits, as well as classes for representing solutions to problems where we are searching for an optimal vector of integers or reals. For each of the built-in representations, the library provides the most common mutation operators for generating random neighbors of candidate solutions, as well as common crossover operators for use with evolutionary algorithms. Additionally, the library provides extensive support for permutation optimization problems, including implementations of many different mutation operators for permutations, and utilizing the efficiently implemented Permutation class of the JavaPermutationTools (JPT) library. Chips-n-Salsa is customizable, making extensive use of Java's generic types, enabling using the library to optimize other types of representations beyond what is provided in the library. It is hybridizable, providing support for integrating multiple forms of local search (e.g., using a hill climber on a solution generated by simulated annealing), creating hybrid mutation operators (e.g., local search using multiple mutation operators), as well as support for running more than one type of search for the same problem concurrently using multiple threads as a form of algorithm portfolio. Chips-n-Salsa is iterative, with support for multistart metaheuristics, including implementations of several restart schedules for varying the run lengths across the restarts. It also supports parallel execution of multiple instances of the same, or different, stochastic local search algorithms for an instance of a problem to accelerate the search process. The library supports self-adaptive search in a variety of ways, such as including implementations of adaptive annealing schedules for simulated annealing, such as the Modified Lam schedule, implementations of the simpler annealing schedules but which self-tune the initial temperature and other parameters, and restart schedules that adapt to run length.

There is a newer version: 7.0.1

Show newest version

/*
 * Chips-n-Salsa: A library of parallel self-adaptive local search algorithms.
 * Copyright (C) 2002-2022 Vincent A. Cicirello
 *
 * This file is part of Chips-n-Salsa (https://chips-n-salsa.cicirello.org/).
 *
 * Chips-n-Salsa 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.
 *
 * Chips-n-Salsa 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 this program.  If not, see .
 */

package org.cicirello.search.evo;

/**
 * This class implements Boltzmann selection using Stochastic Universal Sampling (SUS). Boltzmann
 * selection is similar to a fitness proportional selection, except instead of a population member
 * being weighted by its fitness f in the randomized selection process, Boltzmann selection weights
 * it by e^f/T, where f is the fitness of the individual, and T is a temperature
 * parameter, much like that of simulated annealing. T typically decreases over the run of the
 * evolutionary algorithm.
 *
 * This implementation supports a constant temperature T, as well as two cooling schedules:
 * linear cooling and exponential cooling. In both cases, at the start of the evolutionary
 * algorithm, the temperature T is initialized to a t0. In linear cooling, at the end of each
 * generation, T is updated according to: T = T - r. In exponential cooling, at the end of each
 * generation, T is updated according to: T = r * T. In both cases, if T ever falls below some tMin,
 * it is reset to tMin.
 *
 * 
Unlike many other fitness proportional related selection operators, Boltzmann selection,
 * including this SUS version, is applicable even if fitness values can be negative.
 *
 * 
SUS and this Boltzmann SUS are similar to fitness proportional selection and a variation of
 * fitness proportional selection biasing selection by the Boltzmann distribution. However, whereas
 * fitness proportional selection is like spinning a carnival wheel with a single pointer M times to
 * select M members of the population, SUS instead is like spinning a carnival wheel that has M
 * equidistant pointers a single time to select all M simultaneously. One statistical consequence of
 * this is that it reduces the variance of the selected copies of population members as compared to
 * fitness proportional selection. Another consequence is that SUS is typically much faster since
 * only a single random floating point number is needed per generation, compared to M random
 * floating-point numbers for fitness proportional selection. However, SUS then must randomize the
 * ordering of the population to avoid all of the copies of a single population member from being in
 * sequence so that parent assignment is random, whereas fitness proportional selection has this
 * property built in.
 *
 * 
The runtime to select M population members from a population of size N is O(N + M), which
 * includes the need to generate only a single random double, and O(M) ints.
 *
 * For the basic version of Boltzmann selection, see the {@link BoltzmannSelection} class. And
 * for the standard version of SUS, see the {@link StochasticUniversalSampling} class.
 *
 * @author Vincent A. Cicirello, https://www.cicirello.org/
 */
public final class BoltzmannStochasticUniversalSampling extends BiasedStochasticUniversalSampling {

  private final BoltzmannBiasFunction boltzmann;

  /**
   * Construct a Boltzmann selection operator with a constant temperature.
   *
   * @param t The temperature, which must be positive.
   * @throws IllegalArgumentException if t is not positive
   */
  public BoltzmannStochasticUniversalSampling(double t) {
    this(new ConstantBoltzmannBiasFunction(t));
    if (t <= 0.0) throw new IllegalArgumentException("The temperature must be positive.");
  }

  /**
   * Construct a Boltzmann selection operator, with either a linear cooling schedule or an
   * exponential cooling schedule.
   *
   * @param t0 The initial temperature, which must be positive.
   * @param tMin The minimum temperature. If an update would decrease temperature below tMin, it is
   *     set to tMin. Must be positive and no greater than t0.
   * @param r The update value, which must be positive for linear cooling, and must be in (0.0, 1.0)
   *     for exponential cooling.
   * @param linearCooling If true, uses a linear cooling schedule, and if false it uses an
   *     exponential cooling schedule.
   * @throws IllegalArgumentException if tMin is not positive or if t0 is less than tMin or if
   *     linear cooling with non-positive r or if exponential cooling with r not in (0,0, 1.0).
   */
  public BoltzmannStochasticUniversalSampling(
      double t0, double tMin, double r, boolean linearCooling) {
    this(
        linearCooling
            ? new LinearCoolingBiasFunction(t0, r, tMin)
            : new ExponentialCoolingBiasFunction(t0, r, tMin));
    if (tMin <= 0.0) throw new IllegalArgumentException("Minimum temperature must be positive.");
    if (t0 < tMin)
      throw new IllegalArgumentException(
          "Minimum temperature must be no greater than initial temperature.");
    if (r <= 0.0) throw new IllegalArgumentException("r must be positive");
    if (!linearCooling && r >= 1.0)
      throw new IllegalArgumentException("For exponential cooling, r must be positive.");
  }

  private BoltzmannStochasticUniversalSampling(BoltzmannBiasFunction boltzmann) {
    super(boltzmann);
    this.boltzmann = boltzmann;
  }

  private BoltzmannStochasticUniversalSampling(BoltzmannStochasticUniversalSampling other) {
    this(other.boltzmann.split());
  }

  @Override
  public void init(int generations) {
    boltzmann.init();
  }

  @Override
  public BoltzmannStochasticUniversalSampling split() {
    return new BoltzmannStochasticUniversalSampling(this);
  }

  @Override
  final double[] computeWeightRunningSum(PopulationFitnessVector.Integer fitnesses) {
    double[] result = super.computeWeightRunningSum(fitnesses);
    boltzmann.update();
    return result;
  }

  @Override
  final double[] computeWeightRunningSum(PopulationFitnessVector.Double fitnesses) {
    double[] result = super.computeWeightRunningSum(fitnesses);
    boltzmann.update();
    return result;
  }
}