Reasoning about interactive systems with stochastic models

Several techniques for specification exist to capture certain aspects of user behaviour, with the goal of reasoning about the usability of the system and other human-factors related issues. One such approach is to encode a set of assumptions about user behaviour in a user model. A difficulty with this approach is that human behaviour is inherently nondeterministic; humans make errors, perform unexpected actions, and, taken individually, both the occurrence of errors and response times can be unpredictable. Such factors, however can be expected to follow probability distributions, and so an interesting possibility is to apply stochasticor probabilistictec hniques that allow the modelling of uncertainty in user models. Recently, a number of process algebra based approaches to specifying stochastic systems have been proposed and in this paper we examine the possibility of applying these stochastic modelling techniques to reasoning about performance aspects of interactive systems.