Weighted popular matchings

  • Authors:
  • Julián Mestre

  • Affiliations:
  • Department of Computer Science, University of Maryland, College Park, MD

  • Venue:
  • ICALP'06 Proceedings of the 33rd international conference on Automata, Languages and Programming - Volume Part I
  • Year:
  • 2006

Quantified Score

Hi-index 0.00

Visualization

Abstract

We study the problem of assigning applicants to jobs. Each applicant has a weight and provides a preference list, which may contain ties, ranking a subset of the jobs. An applicant x may prefer one matching over the other (or be indifferent between them, in case of a tie) based on the jobs x gets in the two matchings and x's personal preference. A matching M is popular if there is no other matching M′ such that the weight of the applicants who prefer M′ over M exceeds the weight of those who prefer M over M′. We present two algorithms to find a popular matching, or in case none exists, to establish so. For the case of strict preferences we develop an O(n+m) time algorithm. When ties are allowed a more involved algorithm solves the problem in $O(\min(k \sqrt{n}, n) m)$ time, where k is the number of distinct weights the applicants are given.