Skip to main navigation Skip to search Skip to main content

A Stochastic Optimization Algorithm Minimizing Expected Flow Times on Uniform Processors

  • Ashok K. Agrawala
  • , Satish K. Tripathi
  • , Edward G. Coffman
  • , Michael R. Garey
  • University of Maryland, College Park
  • Nokia

Research output: Contribution to journalArticlepeer-review

43 Scopus citations

Abstract

Consider a set of processors P1, • • •, Pmdiffering only in speed, and a set of jobs with exponentially distributed execution times. The rate parameter for the ith processor is given by, 1 < i < m, where we assume the processors are ordered so that μ1 • • • >μ 2.> μm. The problem is to sequence the jobs nonpreemptively so as to minimize expected total flow time (sum of finishing times). We define a threshold rule that assigns a job to a lowest indexed available processor Pj if and only if the number k of waiting jobs satisfies • • + k > (j — 1).+• My Our main result is a proof that this rule minimizes expected total flow time.

Original languageEnglish
Pages (from-to)351-356
Number of pages6
JournalIEEE Transactions on Computers
VolumeC-33
Issue number4
DOIs
StatePublished - Apr 1984

Keywords

  • Mean flow time minimization
  • routing problems
  • stochastic optimization
  • stochastic scheduling
  • uniform processor systems

Fingerprint

Dive into the research topics of 'A Stochastic Optimization Algorithm Minimizing Expected Flow Times on Uniform Processors'. Together they form a unique fingerprint.

Cite this