Behavior of task scheduling algorithms with probabilistic module execution times

M. Kholief, I. Awad, M. El-Derini, M. Nagi, H. Ali

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Static task scheduling in distributed computing systems is a very complex problem and known to be NP-hard. This problem is even harder when the module execution times become probabilistic. In this paper we study the effect of probabilistic module execution times on the performance of task-scheduling algorithms. We show that in static task scheduling, for probabilistic module execution times, and in the existence of some factors there is no need to use an expensive task-scheduling algorithm. Given any two static task-scheduling algorithms that use deterministic module execution times in assigning task modules to the distributed system, the performance of these two algorithms will not remain the same when these module execution times become probabilistic rather than deterministic. We also study the effects of some factors an our results.

Original languageEnglish (US)
Title of host publicationProceedings - 3rd IEEE Symposium on Computers and Communications, ISCC 1998
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages647-651
Number of pages5
ISBN (Print)0818685387, 9780818685385
DOIs
StatePublished - 1998
Externally publishedYes
Event3rd IEEE Symposium on Computers and Communications, ISCC 1998 - Athens, Greece
Duration: Jun 30 1998Jul 2 1998

Publication series

NameProceedings - 3rd IEEE Symposium on Computers and Communications, ISCC 1998

Other

Other3rd IEEE Symposium on Computers and Communications, ISCC 1998
Country/TerritoryGreece
CityAthens
Period6/30/987/2/98

ASJC Scopus subject areas

  • Computer Networks and Communications
  • Safety, Risk, Reliability and Quality

Fingerprint

Dive into the research topics of 'Behavior of task scheduling algorithms with probabilistic module execution times'. Together they form a unique fingerprint.

Cite this