org.cicirello.search.problems.PermutationInAHaystack 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) 2002-2021 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.problems;

import org.cicirello.permutations.Permutation;
import org.cicirello.permutations.distance.PermutationDistanceMeasurer;

/**
 * The Permutation in a Haystack is a family of optimization problems that can be parameterized to
 * the various types of permutation problem (e.g., absolute versus relative positioning). The
 * Permutation in a Haystack problem uses permutation distance metrics to specify search landscape
 * topology, providing an easy means of studying the behavior of search operators on a wide variety
 * of permutation landscapes.
 *
 * The Permutation in a Haystack Problem, Haystack(δ, n), is defined as follows. Find the
 * permutation p such that, p = argmin_p' δ(p', p_n), where p_n
 * = [0, 1, ..., (n-1)]. The p_n is called the target permutation, and is just the
 * permutation of the first n non-negative integers in increasing order. The p_n is our
 * figurative needle for which we are searching our haystack. It is also the optimal solution to the
 * problem, so in this way the search problem has a known optimal. The δ is a measure of the
 * distance for permutations. There are many measures of permutation distance available in the
 * literature. Some focus on exact positions of elements in the permutation, others focus on
 * relative ordering of permutation elements, others focus on element precedences, etc. In this way,
 * the choice of δ enables you to control the search space topology. The {@link
 * org.cicirello.permutations.distance} package includes implementations of many permutation
 * distance measures. The class includes a constructor that uses the target permutation as defined
 * above, as well as an additional constructor that enables you to specify a different target.
 *
 * The Permutation in a Haystack Problem was introduced in the following paper:

 * V.A. Cicirello, "The Permutation in a Haystack Problem and the Calculus of Search Landscapes,"
 * IEEE Transactions on Evolutionary Computation, 20(3):434-446, June 2016.
 *
 * @author Vincent A. Cicirello, https://www.cicirello.org/
 */
public final class PermutationInAHaystack implements IntegerCostOptimizationProblem {

  private final PermutationDistanceMeasurer distance;
  private final Permutation target;

  /**
   * Constructs an instance of the Permutation in a Haystack problem, for a given distance measure.
   * The target permutation, the figurative needle for which we are searching the haystack, is set
   * to the following permutation: [0, 1, ..., (n-1)]. That is, the known optimal solution to the
   * problem is just the permutation of the first n integers in increasing order.
   *
   * @param distance A permutation distance measure,
   * @param n The length of the target permutation.
   */
  public PermutationInAHaystack(PermutationDistanceMeasurer distance, int n) {
    this.distance = distance;
    int[] p = new int[n];
    for (int i = 0; i < n; i++) p[i] = i;
    target = new Permutation(p);
  }

  /**
   * Constructs an instance of the Permutation in a Haystack problem, for a given distance measure,
   * and given target permutation.
   *
   * @param distance A permutation distance measure,
   * @param target The target permutation, such that the problem is to find the permutation with
   *     minimum distance to the target. That is, target is our figurative needle for which we are
   *     searching the haystack. It is the known optimal solution to the problem.
   */
  public PermutationInAHaystack(PermutationDistanceMeasurer distance, Permutation target) {
    this.distance = distance;
    this.target = new Permutation(target);
  }

  @Override
  public int cost(Permutation candidate) {
    return distance.distance(candidate, target);
  }

  @Override
  public int minCost() {
    return 0;
  }

  @Override
  public int value(Permutation candidate) {
    return distance.distance(candidate, target);
  }

  @Override
  public boolean isMinCost(int cost) {
    return cost == 0;
  }
}