Bounds and Approximations for a Generalized Measure of Performance in the G/N/1/N Queue.
Abstract
A generalized measure of performance is defined as a weighted combination of the ergodic queue length distribution where the weights are general functions of the system parameters. The paper presents a sequence of upper and lower bounds for this measure of performance in the G/M/1/N queue with FIFO discipline. The bounds are used to derive a sequence of approximations with bounded errors. The upper and lower bounds are shown to converge to their corresponding exact values. The technique used is based on the imbedded Markov chain analysis and considers only subsets of the steady state equations to derive the bounds. Initial computational experience is encouraging and has indicated that the approximations are viable for heavy and medium traffic conditions. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Feb 01, 1978
- Accession Number
- ADA053497
Entities
People
- Mohamed A. Shalaby
- U. Narayan Bhat
Organizations
- Southern Methodist University