Stochastic Process Algebra -- A formal approach to performance modelling
Speaker: Jane Hillston
Time: Mon 31st May 1999, at 16.15
Place: Room 2.33
Abstract:
Since their introduction, approximately ten years ago, stochastic
process algebras have gained some acceptance within the performance
community as a high-level modelling paradigm for Markov processes. In
this talk I will motivate the development of stochastic process
algebra languages such as PEPA, and focus on the benefits that the
formal structure of these languages bring to performance modelling.
These benefits, based on equivalence relations and the operational
semantics of the language, have largely been aimed at addressing the
state space explosion problem of the generated Markov processes.
However, there is another problem associated with state-based
modelling techniques -- that of deriving performance measures from the
steady state probability distribution resulting from numerical
analysis of the Markov process. In the later part of the talk I will
discuss recent work with Clark and Gilmore on a companion language for
PEPA which allows performance measures to be formally specified in
terms of the high-level model.