Symbolic Approaches to Probability Distributions in Process Algebra
Chris Tofts · Formal Aspects of Computing · 2000
Abstract. One of the major problems in modelling with process algebras is that of state explosion. With the addition of information describing duration, models become more exposed to growth in the size of the description of their behaviour. We describe a symbolic approach to the description of duration information within a probabilistic calculus that has two benefits: firstly we can describe behaviour without instantiating distributions; secondly we can control the growth in our system representation as a trade against model accuracy. We can compute upper bounds on performance without the inclusion of additional state information in our systems. A large example, from the discrete event simulation literature, is presented to demonstrate the effectiveness of the approach.