eCite Digital Repository
Generalised reward generator for stochastic fluid models
Citation
Samuelson, A and O'Reilly, MM and Bean, NB, Generalised reward generator for stochastic fluid models, Proceedings of the 9th International Conference on Matrix-Analytic Methods in Stochastic Models, 28-30 June 2016, Budapest, Hungary, pp. 27-34. ISBN 9781450321389 (2016) [Refereed Conference Paper]
![]() | PDF 1Mb |
Copyright Statement
Copyright unknown
Abstract
We construct a generalised reward matrix Z(s). which is an extension of the fluid generator Q(s) of a stochastic fluid model (SFM). We classify the generators that are projections of Z(s), including the generator Q(s), and discuss the application of the resulting generators in different contexts.
As one application example, for the case with nonzero mean drift, we derive a new Riccati equation for the key matrix Ψ, which records the probabilities of the first return to the original level.
The Riccati equation has the form Ψ + ΨM-+Ψ = M+-, where parameters M+- and M-+ are block matrices in the matrix M, which records the expected number of visits to the original level, before the unbounded fluid drifts to ± ∞.
Finally, we derive the explicit form Ψ = M+- (I + M--)-1.
Item Details
Item Type: | Refereed Conference Paper |
---|---|
Keywords: | Markov chains, stochastic fluid models, tandem |
Research Division: | Mathematical Sciences |
Research Group: | Statistics |
Research Field: | Stochastic analysis and modelling |
Objective Division: | Expanding Knowledge |
Objective Group: | Expanding knowledge |
Objective Field: | Expanding knowledge in the mathematical sciences |
UTAS Author: | Samuelson, A (Ms Aviva Samuelson) |
UTAS Author: | O'Reilly, MM (Associate Professor Malgorzata O'Reilly) |
ID Code: | 112933 |
Year Published: | 2016 |
Funding Support: | Australian Research Council (LP140100152) |
Deposited By: | Mathematics and Physics |
Deposited On: | 2016-12-05 |
Last Modified: | 2018-03-20 |
Downloads: | 58 View Download Statistics |
Repository Staff Only: item control page