A logarithmic reduction algorithm for quasi-birth-death processes
- 1 June 1993
- journal article
- research article
- Published by Cambridge University Press (CUP) in Journal of Applied Probability
- Vol. 30 (03) , 650-674
- https://doi.org/10.1017/s0021900200044387
Abstract
Quasi-birth-death processes are commonly used Markov chain models in queueing theory, computer performance, teletraffic modeling and other areas. We provide a new, simple algorithm for the matrix-geometric rate matrix. We demonstrate that it has quadratic convergence. We show theoretically and through numerical examples that it converges very fast and provides extremely accurate results even for almost unstable models.Keywords
This publication has 18 references indexed in Scilit:
- Equilibrium distribution of block-structured Markov chains with repeating rowsJournal of Applied Probability, 1990
- Probabilistic interpretations of some duality results for the matrix paradigms in queueing theoryCommunications in Statistics. Stochastic Models, 1990
- A duality theorem for the matrix paradigms in queueing theoryCommunications in Statistics. Stochastic Models, 1990
- Experimental results on matrix-analytical solution techniques–extensions and comparisonsCommunications in Statistics. Stochastic Models, 1989
- An experimental evaluation of the matrix-geometric method for theGI/PH/1 queueCommunications in Statistics. Stochastic Models, 1989
- A note on two matrices occurring in the solution of quasi-birth-and-death processesCommunications in Statistics. Stochastic Models, 1987
- The caudal characteristic curve of queuesAdvances in Applied Probability, 1986
- Algorithmic Analysis of a Multiprogramming-Multiprocessor Computer SystemJournal of the ACM, 1981
- A Duality Theorem for Phase Type QueuesThe Annals of Probability, 1980
- Moment formulas for the Markov renewal branching processAdvances in Applied Probability, 1976