org.jgrapht.alg.matching.GreedyWeightedMatching Maven / Gradle / Ivy
/*
* (C) Copyright 2016-2021, by Dimitrios Michail and Contributors.
*
* JGraphT : a free Java graph-theory library
*
* See the CONTRIBUTORS.md file distributed with this work for additional
* information regarding copyright ownership.
*
* This program and the accompanying materials are made available under the
* terms of the Eclipse Public License 2.0 which is available at
* http://www.eclipse.org/legal/epl-2.0, or the
* GNU Lesser General Public License v2.1 or later
* which is available at
* http://www.gnu.org/licenses/old-licenses/lgpl-2.1-standalone.html.
*
* SPDX-License-Identifier: EPL-2.0 OR LGPL-2.1-or-later
*/
package org.jgrapht.alg.matching;
import org.jgrapht.*;
import org.jgrapht.alg.interfaces.*;
import org.jgrapht.alg.util.*;
import java.util.*;
/**
* The greedy algorithm for computing a maximum weight matching in an arbitrary graph. The algorithm
* runs in $O(m + m \log n)$ where $n$ is the number of vertices and $m$ is the number of edges of
* the graph. This implementation accepts directed and undirected graphs which may contain
* self-loops and multiple (parallel) edges. There is no assumption on the edge weights, i.e. they
* can also be negative or zero.
*
*
* This algorithm can be run in two modes: with and without edge cost normalization. Without
* normalization, the algorithm first orders the edge set in non-increasing order of weights and
* then greedily constructs a maximal cardinality matching out of the edges with positive weight. A
* maximal cardinality matching (not to be confused with maximum cardinality) is a matching that
* cannot be increased in cardinality without removing an edge first. The resulting matching is
* guaranteed to be a $\frac{1}{2}$-Approximation.
* With normalization, the edges are sorted in non-increasing order of their normalized costs
* $\frac{c(u,v)}{d(u)+d(v)}$ instead, after which the algorithm proceeds in the same manner. Here,
* $c(u,v)$ is the cost of edge $(u,v)$, and $d(u)$ resp $d(v)$ are the degrees of vertices $u$ resp
* $v$. Running this algorithm in normalized mode often (but not always!) produces a better result
* than running the algorithm without normalization. Note however that the normalized version
* does NOT produce a $\frac{1}{2}$-approximation. See this
* proof for details. The runtime complexity remains the same, independent of whether
* normalization is used.
*
*
* For more information about approximation algorithms for the maximum weight matching problem in
* arbitrary graphs see:
*
* - R. Preis, Linear Time $\frac{1}{2}$-Approximation Algorithm for Maximum Weighted Matching in
* General Graphs. Symposium on Theoretical Aspects of Computer Science, 259-269, 1999.
* - D.E. Drake, S. Hougardy, A Simple Approximation Algorithm for the Weighted Matching Problem,
* Information Processing Letters 85, 211-213, 2003.
*
*
* @see PathGrowingWeightedMatching
*
* @param the graph vertex type
* @param the graph edge type
*
* @author Dimitrios Michail
*/
public class GreedyWeightedMatching
implements
MatchingAlgorithm
{
private final Graph graph;
private final Comparator comparator;
private final boolean normalizeEdgeCosts;
/**
* Create and execute a new instance of the greedy maximum weight matching algorithm. Floating
* point values are compared using {@link #DEFAULT_EPSILON} tolerance.
*
* @param graph the input graph
* @param normalizeEdgeCosts boolean indicating whether edge normalization has to be used.
*/
public GreedyWeightedMatching(Graph graph, boolean normalizeEdgeCosts)
{
this(graph, normalizeEdgeCosts, DEFAULT_EPSILON);
}
/**
* Create and execute a new instance of the greedy maximum weight matching algorithm.
*
* @param graph the input graph
* @param normalizeEdgeCosts boolean indicating whether edge normalization has to be used.
* @param epsilon tolerance when comparing floating point values
*/
public GreedyWeightedMatching(Graph graph, boolean normalizeEdgeCosts, double epsilon)
{
if (graph == null) {
throw new IllegalArgumentException("Input graph cannot be null");
}
this.graph = graph;
this.comparator = new ToleranceDoubleComparator(epsilon);
this.normalizeEdgeCosts = normalizeEdgeCosts;
}
/**
* Get a matching that is a $\frac{1}{2}$-approximation of the maximum weighted matching.
*
* @return a matching
*/
@Override
public Matching getMatching()
{
// sort edges in non-decreasing order of weight
// (the lambda uses e1 and e2 in the reverse order on purpose)
List allEdges = new ArrayList<>(graph.edgeSet());
if (normalizeEdgeCosts) {
allEdges.sort((e1, e2) -> {
double degreeE1 = graph.degreeOf(graph.getEdgeSource(e1))
+ graph.degreeOf(graph.getEdgeTarget(e1));
double degreeE2 = graph.degreeOf(graph.getEdgeSource(e2))
+ graph.degreeOf(graph.getEdgeTarget(e2));
return comparator
.compare(
graph.getEdgeWeight(e2) / degreeE2, graph.getEdgeWeight(e1) / degreeE1);
});
} else {
allEdges
.sort(
(e1, e2) -> comparator
.compare(graph.getEdgeWeight(e2), graph.getEdgeWeight(e1)));
}
double matchingWeight = 0d;
Set matching = new HashSet<>();
Set matchedVertices = new HashSet<>();
// find maximal matching
for (E e : allEdges) {
double edgeWeight = graph.getEdgeWeight(e);
V s = graph.getEdgeSource(e);
V t = graph.getEdgeTarget(e);
if (!s.equals(t) && comparator.compare(edgeWeight, 0d) > 0) {
if (!matchedVertices.contains(s) && !matchedVertices.contains(t)) {
matching.add(e);
matchedVertices.add(s);
matchedVertices.add(t);
matchingWeight += edgeWeight;
}
}
}
// return matching
return new MatchingImpl<>(graph, matching, matchingWeight);
}
}