001    package net.sf.cpsolver.ifs.example.csp;
002    
003    import java.util.Iterator;
004    import java.util.Random;
005    
006    import net.sf.cpsolver.ifs.model.Constraint;
007    import net.sf.cpsolver.ifs.model.Model;
008    
009    /**
010     * Random Binary CSP with uniform distribution. <br>
011     * <br>
012     * A random CSP is defined by a four-tuple (n, d, p1, p2), where n denotes the
013     * number of variables and d denotes the domain size of each variable, p1 and p2
014     * are two probabilities. They are used to generate randomly the binary
015     * constraints among the variables. p1 represents the probability that a
016     * constraint exists between two different variables and p2 represents the
017     * probability that a pair of values in the domains of two variables connected
018     * by a constraint are incompatible. <br>
019     * <br>
020     * We use a so called model B of Random CSP (n, d, n1, n2) where n1 =
021     * p1*n*(n-1)/2 pairs of variables are randomly and uniformly selected and
022     * binary constraints are posted between them. For each constraint, n2 = p1*d^2
023     * randomly and uniformly selected pairs of values are picked as incompatible.
024     * 
025     * @version IFS 1.2 (Iterative Forward Search)<br>
026     *          Copyright (C) 2006 - 2010 Tomas Muller<br>
027     *          <a href="mailto:muller@unitime.org">muller@unitime.org</a><br>
028     *          <a href="http://muller.unitime.org">http://muller.unitime.org</a><br>
029     * <br>
030     *          This library is free software; you can redistribute it and/or modify
031     *          it under the terms of the GNU Lesser General Public License as
032     *          published by the Free Software Foundation; either version 3 of the
033     *          License, or (at your option) any later version. <br>
034     * <br>
035     *          This library is distributed in the hope that it will be useful, but
036     *          WITHOUT ANY WARRANTY; without even the implied warranty of
037     *          MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
038     *          Lesser General Public License for more details. <br>
039     * <br>
040     *          You should have received a copy of the GNU Lesser General Public
041     *          License along with this library; if not see
042     *          <a href='http://www.gnu.org/licenses/'>http://www.gnu.org/licenses/</a>.
043     */
044    public class CSPModel extends Model<CSPVariable, CSPValue> {
045    
046        /**
047         * Constructor
048         * 
049         * @param nrVariables
050         *            number of variables in the problem
051         * @param nrValues
052         *            number of values of each variable
053         * @param nrConstraints
054         *            number of constraints in the problem
055         * @param nrCompatiblePairs
056         *            number of compatible pairs of values for every constraint
057         * @param seed
058         *            seed for random number generator (use
059         *            {@link System#currentTimeMillis} if not bother)
060         */
061        public CSPModel(int nrVariables, int nrValues, int nrConstraints, int nrCompatiblePairs, long seed) {
062            generate(nrVariables, nrValues, nrConstraints, nrCompatiblePairs, seed);
063        }
064    
065        public CSPModel() {
066        }
067    
068        private void swap(CSPVariable[][] allPairs, int first, int second) {
069            CSPVariable[] a = allPairs[first];
070            allPairs[first] = allPairs[second];
071            allPairs[second] = a;
072        }
073    
074        private void buildBinaryConstraintGraph(Random rnd) {
075            int numberOfAllPairs = variables().size() * (variables().size() - 1) / 2;
076            CSPVariable[][] allPairs = new CSPVariable[numberOfAllPairs][];
077            int idx = 0;
078            for (CSPVariable v1 : variables()) {
079                for (CSPVariable v2 : variables()) {
080                    if (v1.getId() >= v2.getId())
081                        continue;
082                    allPairs[idx++] = new CSPVariable[] { v1, v2 };
083                }
084            }
085            idx = 0;
086            for (Iterator<Constraint<CSPVariable, CSPValue>> i = constraints().iterator(); i.hasNext();) {
087                CSPBinaryConstraint c = (CSPBinaryConstraint) i.next();
088                swap(allPairs, idx, idx + (int) (rnd.nextDouble() * (numberOfAllPairs - idx)));
089                c.addVariable(allPairs[idx][0]);
090                c.addVariable(allPairs[idx][1]);
091                c.init(rnd);
092                idx++;
093            }
094        }
095    
096        private void generate(int nrVariables, int nrValues, int nrConstraints, int nrCompatiblePairs, long seed) {
097            Random rnd = new Random(seed);
098    
099            for (int i = 0; i < nrVariables; i++) {
100                CSPVariable var = new CSPVariable(i + 1, nrValues);
101                addVariable(var);
102            }
103    
104            for (int i = 0; i < nrConstraints; i++) {
105                CSPBinaryConstraint c = new CSPBinaryConstraint(i + 1, nrCompatiblePairs);
106                addConstraint(c);
107            }
108    
109            buildBinaryConstraintGraph(rnd);
110        }
111    }