Exact algorithms and APX-hardness results for geometric packing and covering problems

  • Authors:
  • Timothy M. Chan;Elyot Grant

  • Affiliations:
  • David R. Cheriton School of Computer Science, University of Waterloo, Canada;Computer Science and Artificial Intelligence Laboratory, Massachusetts Institute of Technology, United States

  • Venue:
  • Computational Geometry: Theory and Applications
  • Year:
  • 2014

Quantified Score

Hi-index 0.00

Visualization

Abstract

We study several geometric set cover and set packing problems involving configurations of points and geometric objects in Euclidean space. We show that it is APX-hard to compute a minimum cover of a set of points in the plane by a family of axis-aligned fat rectangles, even when each rectangle is an @e-perturbed copy of a single unit square. We extend this result to several other classes of objects including almost-circular ellipses, axis-aligned slabs, downward shadows of line segments, downward shadows of graphs of cubic functions, fat semi-infinite wedges, 3-dimensional unit balls, and axis-aligned cubes, as well as some related hitting set problems. We also prove the APX-hardness of a related family of discrete set packing problems. Our hardness results are all proven by encoding a highly structured minimum vertex cover problem which we believe may be of independent interest. In contrast, we give a polynomial-time dynamic programming algorithm for geometric set cover where the objects are pseudodisks containing the origin or are downward shadows of pairwise 2-intersecting x-monotone curves. Our algorithm extends to the weighted case where a minimum-cost cover is required. We give similar algorithms for several related hitting set and discrete packing problems.