We study the following generalization of the maximum matching problem in general graphs: Given a simple non-directed graph G=(V,E) and a partition of the edges into k classes (i.e. E=E 1∩⋯∩E k ), we would like to compute a matching M on G of maximum cardinality or profit, such that |M∩E j |≤wbj for every class Eb j. Such problems were first studied in the context of network design in [17]. We study the problem from a linear programming point of view: We provide a polynomial time -approximation algorithm for the weighted case, matching the integrality gap of the natural LP formulation of the problem. For this, we use and adapt the technique of approximate convex decompositions [19] together with a different analysis and a polyhedral characterization of the natural linear program to derive our result. This improves over the existing, but with additive violation of the color bounds, approximation algorithm [14]. © 2014 Springer-Verlag Berlin Heidelberg.
CITATION STYLE
Stamoulis, G. (2014). Approximation algorithms for bounded color matchings via convex decompositions. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 8635 LNCS, pp. 625–636). Springer Verlag. https://doi.org/10.1007/978-3-662-44465-8_53
Mendeley helps you to discover research relevant for your work.