On the complexity of computing the measure of ∪[ai,bi]

  • Authors:
  • Michael L. Fredman;Bruce Weide

  • Affiliations:
  • The Univ. of California, San Diego;Carnegie-Mellon Univ., Pittsburgh, PA

  • Venue:
  • Communications of the ACM
  • Year:
  • 1978

Quantified Score

Hi-index 48.24

Visualization

Abstract

The decision tree complexity of computing the measure of the union of n (possibly overlapping) intervals is shown to be &OHgr;(n log n), even if comparisons between linear functions of the interval endpoints are allowed. The existence of an &OHgr;(n log n) lower bound to determine whether any two of n real numbers are within ∈ of each other is also demonstrated. These problems provide an excellent opportunity for discussing the effects of the computational model on the ease of analysis and on the results produced.