Difference Equations and the Optimal Control of Single Server Queueing Systems

Abstract

This report demonstrates the use of difference equations in solving optimal control problems in single server queueing systems. One obtains the discounted or relative cost function associated with a specific stationary policy by solving an appropriate system of difference equations. The policy improvement algorithm is applied parametrically leading to a characterization of the cost function satisfying the functional equation of optimality. If this cost function satisfies an appropriate sufficient condition, the associated stationary policy is optimal. The method of solution is illustrated by solving three queueing optimization problems. These problems include optimal control of the M/G/1 queue with intermittent service, a bulk queueing version of this same problem, and control of the M/M/1 queue with selection of running speed. All of these problems have been investigated by other authors. Results in this report believed to be new include a complete characterization of optimal policies for the optimal control of the M/G/1 queue in the discounted case, the extension of the optimal control of the bulk queueing problem from instantaneous to general service, and the determination of an optimal speed selection policy for the M/M/1 queue without solving a sequence of truncated problems.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Jun 01, 1974
Accession Number
ADA371628

Entities

People

  • Frank C. Reed

Organizations

  • Naval Air Weapons Station China Lake

Tags

Communities of Interest

  • C4I
  • Weapons Technologies

DTIC Thesaurus Topics

  • Air Force
  • Algorithms
  • Difference Equations
  • Distribution Functions
  • Equations
  • Jet Propulsion
  • Mathematics
  • Military Research
  • Operations Research
  • Optimization
  • Probability
  • Random Variables
  • Sequences
  • Stochastic Processes
  • Test And Evaluation
  • Universities
  • War Colleges

Readers

  • Mathematical Modeling and Probability Theory.
  • Operations Research