Nonexponential networks of queues: a maximum entropy analysis

  • Authors:
  • Robbe J. Walstra

  • Affiliations:
  • Hewlett-Packard Laboratories, Palo Alto, CA

  • Venue:
  • SIGMETRICS '85 Proceedings of the 1985 ACM SIGMETRICS conference on Measurement and modeling of computer systems
  • Year:
  • 1985

Quantified Score

Hi-index 0.00

Visualization

Abstract

We will propose a new, iterative method for approximately analyzing closed networks of queues with nonexponential service time distributions and FCFS scheduling. Our method is based on the Principle of Maximum Entropy and produces results which, first, are consistent with the fundamental Work Rate Theorem and, second, are exact for separable networks of queues. Considering accuracy and execution time characteristics, our method offers a viable alternative to Marie's homogeneous approximation method.