All Downloads are FREE. Search and download functionalities are using the official Maven repository.

org.chocosolver.solver.constraints.nary.sum.PropScalar Maven / Gradle / Ivy

There is a newer version: 4.10.17
Show newest version
/*
 * This file is part of choco-solver, http://choco-solver.org/
 *
 * Copyright (c) 2020, IMT Atlantique. All rights reserved.
 *
 * Licensed under the BSD 4-clause license.
 *
 * See LICENSE file in the project root for full license information.
 */
package org.chocosolver.solver.constraints.nary.sum;

import org.chocosolver.solver.constraints.Operator;
import org.chocosolver.solver.constraints.nary.clauses.ClauseBuilder;
import org.chocosolver.solver.exception.ContradictionException;
import org.chocosolver.solver.learn.ExplanationForSignedClause;
import org.chocosolver.solver.variables.IntVar;
import org.chocosolver.util.ESat;
import org.chocosolver.util.objects.setDataStructures.iterable.IntIterableRangeSet;

import static org.chocosolver.solver.constraints.Operator.*;

/**
 * A propagator for SUM(x_i*c_i) = b
 * 
* Based on "Bounds Consistency Techniques for Long Linear Constraint"
* W. Harvey and J. Schimpf *

* * @author Charles Prud'homme * @since 18/03/11 */ public class PropScalar extends PropSum { /** * The coefficients */ private final int[] c; /** * Create a scalar product: SUM(x_i*c_i) o b * Variables and coefficients are excepted to be ordered wrt to coefficients: first positive ones then negative ones. * @param variables list of integer variables * @param coeffs list of coefficients * @param pos position of the last positive coefficient * @param o operator * @param b bound to respect. */ public PropScalar(IntVar[] variables, int[] coeffs, int pos, Operator o, int b) { super(variables, pos, o, b); this.c = coeffs; } @Override protected void prepare() { sumLB = sumUB = 0; int i = 0, lb, ub; maxI = 0; for (; i < pos; i++) { // first the positive coefficients lb = vars[i].getLB() * c[i]; ub = vars[i].getUB() * c[i]; sumLB += lb; sumUB += ub; I[i] = (ub - lb); if(maxI < I[i])maxI = I[i]; } for (; i < l; i++) { // then the negative ones lb = vars[i].getUB() * c[i]; ub = vars[i].getLB() * c[i]; sumLB += lb; sumUB += ub; I[i] = (ub - lb); if(maxI < I[i])maxI = I[i]; } } @Override protected void filterOnEq() throws ContradictionException { boolean anychange; int F = b - sumLB; int E = sumUB - b; do { anychange = false; // When explanations are on, no global failure allowed if (model.getSolver().isLearnOff() && F < 0 || E < 0) { fails(); } if (maxI > F || maxI > E) { maxI = 0; int lb, ub, i = 0; // positive coefficients first while (i < pos) { if (I[i] - F > 0) { lb = vars[i].getLB() * c[i]; ub = lb + I[i]; if (vars[i].updateUpperBound(divFloor(F + lb, c[i]), this)) { int nub = vars[i].getUB() * c[i]; E += nub - ub; I[i] = nub - lb; anychange = true; } } if (I[i] - E > 0) { ub = vars[i].getUB() * c[i]; lb = ub - I[i]; if (vars[i].updateLowerBound(divCeil(ub - E, c[i]), this)) { int nlb = vars[i].getLB() * c[i]; F -= nlb - lb; I[i] = ub - nlb; anychange = true; } } if(maxI < I[i])maxI = I[i]; i++; } // then negative ones while (i < l) { if (I[i] - F > 0) { lb = vars[i].getUB() * c[i]; ub = lb + I[i]; if (vars[i].updateLowerBound(divCeil(-F - lb, -c[i]), this)) { int nub = vars[i].getLB() * c[i]; E += nub - ub; I[i] = nub - lb; anychange = true; } } if (I[i] - E > 0) { ub = vars[i].getLB() * c[i]; lb = ub - I[i]; if (vars[i].updateUpperBound(divFloor(-ub + E, -c[i]), this)) { int nlb = vars[i].getUB() * c[i]; F -= nlb - lb; I[i] = ub - nlb; anychange = true; } } if(maxI < I[i])maxI = I[i]; i++; } } if (F <= 0 && E <= 0) { this.setPassive(); return; } } while (anychange); } @Override protected void filterOnLeq() throws ContradictionException { int F = b - sumLB; int E = sumUB - b; // When explanations are on, no global failure allowed if (model.getSolver().isLearnOff() &&F < 0) { fails(); } if (maxI > F) { int lb, ub, i = 0; maxI = 0; // positive coefficients first while (i < pos) { maxI = 0; if (I[i] - F > 0) { lb = vars[i].getLB() * c[i]; ub = lb + I[i]; if (vars[i].updateUpperBound(divFloor(F + lb, c[i]), this)) { int nub = vars[i].getUB() * c[i]; E += nub - ub; I[i] = nub - lb; } } if(maxI < I[i])maxI = I[i]; i++; } // then negative ones while (i < l) { if (I[i] - F > 0) { lb = vars[i].getUB() * c[i]; ub = lb + I[i]; if (vars[i].updateLowerBound(divCeil(-F - lb, -c[i]), this)) { int nub = vars[i].getLB() * c[i]; E += nub - ub; I[i] = nub - lb; } } if(maxI < I[i])maxI = I[i]; i++; } } if (E <= 0) { this.setPassive(); } } @Override protected void filterOnGeq() throws ContradictionException { int F = b - sumLB; int E = sumUB - b; // When explanations are on, no global failure allowed if (model.getSolver().isLearnOff() && E < 0) { fails(); } if (maxI > E) { maxI = 0; int lb, ub, i = 0; // positive coefficients first while (i < pos) { if (I[i] - E > 0) { ub = vars[i].getUB() * c[i]; lb = ub - I[i]; if (vars[i].updateLowerBound(divCeil(ub - E, c[i]), this)) { int nlb = vars[i].getLB() * c[i]; F -= nlb - lb; I[i] = ub - nlb; } } if(maxI < I[i])maxI = I[i]; i++; } // then negative ones while (i < l) { if (I[i] - E > 0) { ub = vars[i].getLB() * c[i]; lb = ub - I[i]; if (vars[i].updateUpperBound(divFloor(-ub + E, -c[i]), this)) { int nlb = vars[i].getUB() * c[i]; F -= nlb - lb; I[i] = ub - nlb; } } if(maxI < I[i])maxI = I[i]; i++; } } if (F <= 0) { this.setPassive(); } } @Override protected void filterOnNeq() throws ContradictionException { int F = b - sumLB; int E = sumUB - b; if (F < 0 || E < 0) { setPassive(); return; } int w = -1; int sum = 0; for (int i = 0; i < l; i++) { if (vars[i].isInstantiated()) { sum += vars[i].getValue() * c[i]; } else if (w == -1) { w = i; } else return; } if (w == -1) { if (sum == b) { this.fails(); } } else if(c[w]!=0 && (b - sum)%c[w]==0){ vars[w].removeValue((b - sum)/c[w], this); } } @Override public ESat isEntailed() { int sumUB = 0, sumLB = 0, i = 0; for (; i < pos; i++) { // first the positive coefficients sumLB += vars[i].getLB() * c[i]; sumUB += vars[i].getUB() * c[i]; } for (; i < l; i++) { // then the negative ones sumLB += vars[i].getUB() * c[i]; sumUB += vars[i].getLB() * c[i]; } return check(sumLB, sumUB); } @Override void doExplain(ExplanationForSignedClause explanation, int p) { IntVar pivot = explanation.readVar(p); IntIterableRangeSet dom_before; // first, compute F and E int sumLB = 0; int sumUB = 0; int i = 0, lb, ub, la = 0, ua = 0, ca = 0, a = 0; for (; i < pos; i++) { // first the positive coefficients dom_before = explanation.readDom(vars[i]); lb = dom_before.min() * c[i]; ub = dom_before.max() * c[i]; if (vars[i] == pivot) { la = dom_before.min(); ua = dom_before.max(); ca = c[i]; a = i; } sumLB += lb; sumUB += ub; } for (; i < l; i++) { // then the negative ones dom_before = explanation.readDom(vars[i]); lb = dom_before.max() * c[i]; ub = dom_before.min() * c[i]; if (vars[i] == pivot) { la = dom_before.min(); ua = dom_before.max(); ca = c[i]; a = i; } sumLB += lb; sumUB += ub; } int F = b - sumLB; int E = sumUB - b; if (explanation.readDom(p).isEmpty()) { doExplainGlobalFailure(explanation, F, E); return; } IntIterableRangeSet domain; int la2 = IntIterableRangeSet.MIN, ua2 = IntIterableRangeSet.MAX; if (ca > 0) { if (!o.equals(GE)) { // ie, LE or EQ ua2 = divFloor(F + la * ca, ca); } if (!o.equals(LE)) { // ie, GE or EQ la2 = divCeil(ca * ua - E, ca); } } else { if (!o.equals(GE)) { // ie, LE or EQ la2 = divCeil(-F - ua * ca, -ca); } if (!o.equals(LE)) { // ie, GE or EQ ua2 = divFloor(-la * ca + E, -ca); } } domain = explanation.empty(); if(la2 <= ua2){ domain.addBetween(la2, ua2); } vars[a].intersectLit(domain, explanation); i = 0; for (; i < pos; i++) { int min = IntIterableRangeSet.MIN; int max = IntIterableRangeSet.MAX; if (vars[i] != pivot) { dom_before = explanation.readDom(vars[i]); if (!o.equals(GE)) { // ie, LE or EQ max = divFloor( F + c[i] * dom_before.min() - ca * (ca > 0 ? (ua2 + 1 - la) : (la2 - 1 - ua)), c[i]); } if (!o.equals(LE)) { // ie, GE or EQ min = divCeil(-E + c[i] * dom_before.max() - ca * (ca > 0 ? la2 - 1 - ua : ua2 + 1 - la), c[i]); } domain = explanation.complement(vars[i]); if(o.equals(EQ)) { assert max+1 <= min-1 : "empty range"; domain.removeBetween(max + 1, min - 1); }else{ domain.retainBetween(min, max); } vars[i].unionLit(domain, explanation); } } for (; i < l; i++) { int min = IntIterableRangeSet.MIN; int max = IntIterableRangeSet.MAX; if (vars[i] != pivot) { dom_before = explanation.readDom(vars[i]); if (!o.equals(GE)) { // ie, LE or EQ min = divCeil( -(F + c[i] * dom_before.max() - ca * (ca > 0 ? ua2 + 1 - la : la2 - 1 - ua)), // done -c[i]); } if (!o.equals(LE)) { // ie, GE or EQ max = divFloor( -(-E + c[i] * dom_before.min() - ca * (ca > 0 ? la2 - 1 - ua : ua2 + 1 - la)) // done , -c[i]); } domain = explanation.complement(vars[i]); if(o.equals(EQ)) { assert max+1 <= min-1 : "empty range"; domain.removeBetween(max + 1, min - 1); }else { domain.retainBetween(min, max); } vars[i].unionLit(domain, explanation); } } } @Override protected void explainGlobal(ExplanationForSignedClause explanation, int F, int E) { assert (F < 0)^(E < 0); IntIterableRangeSet dom_before; IntIterableRangeSet domain; int i = 0; ClauseBuilder ngb = model.getClauseBuilder(); for (; i < l; i++) { int min = IntIterableRangeSet.MIN; int max = IntIterableRangeSet.MAX; dom_before = explanation.readDom(vars[i]); if (F < 0) { // BEWARE // second part of the equation differs from non-global-fail case if(i < pos) { max = divFloor(F + c[i] * dom_before.min(), c[i]); }else{ min = divCeil(-(F + c[i] * dom_before.max()), -c[i]); } }else /*if (E < 0)*/ { // BEWARE // second part of the equation differs from non-global-fail case if(i < pos) { min = divCeil(-E + c[i] * dom_before.max(), c[i]); }else{ max = divFloor(-(-E + c[i] * dom_before.min()), -c[i]); } } domain = explanation.root(vars[i]); domain.retainBetween(min, max); ngb.put(vars[i], domain); int k = 0; for (; k < l; k++) { if (k != i) { min = IntIterableRangeSet.MIN; max = IntIterableRangeSet.MAX; dom_before = explanation.readDom(vars[k]); if (F < 0) { if(k < pos) { min = dom_before.min(); }else{ max = dom_before.max(); } }else /*if (E < 0) */{ if(k < pos) { max = dom_before.max(); }else{ min = dom_before.min(); } } domain = explanation.root(vars[k]); domain.removeBetween(min, max); ngb.put(vars[k], domain); } } ngb.buildNogood(model); if(E == -1 || F == -1)return; // the same nogood will be learned all the time } } @Override public String toString() { StringBuilder linComb = new StringBuilder(20); linComb.append(c[0]).append('.').append(vars[0].getName()); int i = 1; for (; i < pos; i++) { linComb.append(" + ").append(c[i]).append('.').append(vars[i].getName()); } for (; i < l; i++) { linComb.append(" - ").append(-c[i]).append('.').append(vars[i].getName()); } linComb.append(" ").append(o).append(" "); linComb.append(b); return linComb.toString(); } private int divFloor(int a, int b) { // we assume b > 0 if (a >= 0) { return (a / b); } else { return (a - b + 1) / b; } } private int divCeil(int a, int b) { // we assume b > 0 if (a >= 0) { return ((a + b - 1) / b); } else { return a / b; } } @Override protected PropSum opposite(){ return new PropScalar(vars, c, pos, nop(o), b + nb(o)); } }





© 2015 - 2024 Weber Informatics LLC | Privacy Policy