Matching colored points with rectangles

Abstract

Let S be a point set in the plane such that each of its elements is colored either red or blue. A matching of S with rectangles is any set of pairwise-disjoint axis-aligned rectangles such that each rectangle contains exactly two points of S. Such a matching is monochromatic if every rectangle contains points of the same color, and is bichromatic if every rectangle contains points of different colors. In this paper we study the following two problems: 1. Find a maximum monochromatic matching of S with rectangles. 2. Find a maximum bichromatic matching of S with rectangles. For each problem we provide a polynomial-time approximation algorithm that constructs a matching with at least 1/4 of the number of rectangles of an optimal matching. We show that the first problem is NP-hard even if either the matching rectangles are restricted to axis-aligned segments or S is in general position, that is, no two points of S share the same x or y coordinate. We further show that the second problem is also NP-hard, even if S is in general position. These NP-hardness results follow by showing that deciding the existence of a perfect matching is NP-complete in each case. The approximation results are based on a relation of our problem with the problem of finding a maximum independent set in a family of axis-aligned rectangles. With this paper we extend previous ones on matching one-colored points with rectangles and squares, and matching two-colored points with segments. Furthermore, using our techniques, we prove that it is NP-complete to decide a perfect matching with rectangles in the case where all points have the same color, solving an open problem of Bereg, Mutsanas, and Wolff [CGTA (2009)].

0

Turn this paper into a lesson

ArcXiv compiles a structured reading guide from this paper's metadata: plain-English importance, contributions, prerequisite concepts, which sections to read first, flashcards, and a quiz. Grounded in the abstract, never invented.

Discussion (0)

Sign in to join the discussion.

Loading comments…