Analysing Process Models Quantitatively.

Abstract

Over the years, there has been much interest in modelling processes. Processes include those associated with the development of software and those business processes that make use of software systems. Recent research in Systems Engineering for Business Process Change highlights the importance of modelling business processes in order to evolve and maintain the legacy systems that support those processes. Business processes are typically described with static (diagrammatic) models. This paper illustrates how quantitative techniques can facilitate analysis of such models. This is illustrated with reference to the process modelling notation Role Activity Diagrams (RADs). An example process, taken from an investigation of the bidding process of a large telecommunications systems supplier, is used to show how a quantitative approach can be used to highlight features in RADs that are useful to the process modeller. We show how simple measures reveal high levels of role coupling and discrepancies between different perspectives. Since the models are non-trivial — there are 101 roles and almost 300 activities — we argue that quantitative analysis can be a useful adjunct for the modeller.