DSpace About DSpace Software

DSpace Biblioteca Universidad de Talca (v1.5.2) >
Dirección de Investigación >
Artículos en publicaciones ISI - Universidad de Talca >

Please use this identifier to cite or link to this item: http://dspace.utalca.cl/handle/1950/8814

Title: Stronger Quickheaps
Authors: Navarro, G.
Paredes, R.
Poblete, P.V.
Sanders, P.
Keywords: Priority queues
randomized data structures
amortized analysis
Issue Date: Jun-2011
Citation: NTERNATIONAL JOURNAL OF FOUNDATIONS OF COMPUTER SCIENCE Volume: 22 Issue: 4 Special Issue: SI Pages: 945-969
Abstract: The Quickheap (OH) is a recent data structure for implementing priority queues which has proved to be simple and efficient in practice. It has also been shown to offer logarithmic expected amortized complexity for all of its operations. Yet, this complexity holds only when keys inserted and deleted are uniformly distributed over the current set of keys. This assumption is in many cases difficult to verify, and does not hold in some important applications such as implementing some minimum spanning tree algorithms using priority queues. In this paper we introduce an elegant model called a Leftmost Skeleton Tree (LST) that reveals the connection between QIIs and randomized binary search trees, and allows us to define Randomized QHs. We prove that these offer logarithmic expected amortized complexity for all operations regardless of the input distribution. We also use LSTs in connection to alpha-balanced trees to achieve a practical alpha-Balanced QH that offers worst-case amortized logarithmic time bounds for all the operations. Both variants are much more robust than the original QIIs. We show experimentally that randomized Obis behave almost as efficiently as QHs on random inputs, and that they retain their good performance on inputs where that of QHs degrades.
Description: Paredes, R (Paredes, Rodrigo). Univ Talca, Dept Ciencias Computac, Talca, Chile
URI: http://dspace.utalca.cl/handle/1950/8814
ISSN: 0129-0541
Appears in Collections:Artículos en publicaciones ISI - Universidad de Talca

Files in This Item:

File Description SizeFormat

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.


Valid XHTML 1.0! DSpace Software Copyright © 2002-2009  The DSpace Foundation - Feedback