Generalized Semi-Markovian Process Algebra

We present GSMPA, a stochastically timed process algebra that enhances the expressiveness of the stochastically timed process algebra EMPA by allowing arbitrarily distributed durations to be directly modeled through the adoption of the mechanism of action identification as well as the adoption of the preselection policy for alternative actions. GSMPA is introduced together with the definition of its integrated semantics, from which we derive by projection its functional semantics and its performance semantics. The performance semantic models of GSMPA terms are given by extended generalized semi-Markov processes (EGSMPs), an extension, we have introduced, of the well-known generalized semi-Markov processes (GSMPs) which allows us both to deal with the characteristics of GSMPA terms and to exploit already established results concerning the solution of GSMPs based on the notion of insensitivity. Finally we introduce several reduction methods for EGSMPs aimed at obtaining an insensitive GSMP and a notion of equivalence for EGSMPs. 1. University of Bologna, Department of Computer Science, Mura Anteo Zamboni 7, 40127 Bologna, Italy. E-mail: fbravetti, bernardo, [email protected]