A Langevin Interpretation of PEPA Models

In this paper we examine a Langevin interpretation of the stochastic process algebra PEPA. We show how previous work on chemical systems yielding sets of stochastic differential equations (SDEs) can be adapted to the domain of computer systems. Two simple examples are then examined. Their experimental results show a good match between traditional Markovian interpretation of PEPA and the SDE interpretation introduced here. It also raises the problem of boundary conditions which is briefly discussed and for which we propose a solution.