An Experimental Evaluation of Hybrid Data Structures for Searching

  • Authors:
  • Maureen Korda;Rajeev Raman

  • Affiliations:
  • -;-

  • Venue:
  • WAE '99 Proceedings of the 3rd International Workshop on Algorithm Engineering
  • Year:
  • 1999

Quantified Score

Hi-index 0.00

Visualization

Abstract

A common paradigm in data structures is to combine two different kinds of data structures into one, yielding a hybrid data structure with improved resource bounds. We perform an experimental evaluation of hybrid data structures in the context of maintaining a dynamic ordered set whose items have integer or floating-point keys. Among other things we demonstrate clear speedups over library implementations of search trees, both for predecessor queries and updates. Our implementations use very little extra memory compared to search trees, and are also quite generic.