Stochastic Process Algebras for Quantitative Analysis
Jane Hillston
University of Edinburgh (UK)
Abstract: Process algebras were originally conceived to provide models of the
functional behaviour of concurrent systems, allowing the modeller to
study the correctness properties of such systems. However, simple
additions to a process algebra can make it suitable also to support
the analysis of non-functional properties such
as reliability, availability and performance.
In this course we will consider stochastic process algebras in which
the algebra is enriched with activities which have an
associated stochastic duration. We will demonstrate how such a
language can be used for performance modelling of computer
systems and more general forms of analysis of dynamic behaviour.
Moreover, we will show how the compositional structure of the process
algebra can be used to help in the dynamic analysis. Being discrete
state models, stochastic process algebra models are prone to the
problem of state space explosion but recent results on fluid
semantics allow automated approximation of extremely large systems
using continuous state representations. This approach will also be
explained and demonstrated.
Throughout the course the concepts will be illustrated by examples.
-
In the document below you will find the recommended exam projects for this course. I would expect that most of you will be able to find a suitable small project from the projects discussed in the document. However, if there is something else that you would like to undertake as a project, and you think you can do it within the timeframe (deadline 31st May) then please send me an email and we can discuss it.
- Exam options Associated papers:
If you intend to take the exam, please let me know by sending a brief email letting me know by Tuesday 30th April. This will give me an idea of how many submissions to expect and plan time for marking accordingly.
-
Slides:
- Lecture 1: Introduction
- Lecture 2: Stochastic Process Algebras
- Lecture 3: Case Studies
- Lecture 4: Tackling State Space Explosion
- Lecture 5: Collective Dynamics
- Lecture 6: Modelling Biological Processes
- Lecture 7: Modelling Hybrid Systems
-
Example models:
- PEPA model of the WEB1 example
- PEPA model of the WEB2 example
- PEPA model of the WEB4 example
- PEPA model of the PC LAN example with four nodes discussed in Lecture 1
- PEPA model of the PC LAN example with six nodes discussed in Lecture 1
- PEPA model of the web service composition discussed in Lecture 3
People who do not have a background in quantitative modelling, and particularly continuous time Markov chains might find the following lecture notes, from an undergraduate course of Performance Modelling that I teach, useful for providing background to help understand the BISS course.