AESOP home


PROFORMA: Product Forms for Markovian Analysis

Prof. Peter Harrison
Dr Ashok Argent-Katwala
Dr Maria Vigliotti
EPSRC project GR/S45140/01
Started in November 2003
Completed in October 2004
Funded value

The project builds upon the fundamental theorem called RCAT (Reversed Compound Agent Theorem) which facilitates construction of the equilibrium state probabilities of certain Markov chains in product-form -- without requiring the solution of balance equations. The deliverables of the project will include a new classification of Markov chains with product-form in terms of simple conditions relating to the interactions between components; algorithms to detect such chains and provide their steady state solution; conditions delimiting the boundaries of these classes; and, possibly, conditions for ergodicity. From these it should be possible to develop a tool-kit for providing new product-forms from specifications, whilst unifying and essentially subsuming the majority of existing results.