In: Proc. of 6th International Workshop on Petri Nets and Performance Models - PNPM '95, Durham, N. Carolina, USA, pages 113-122. IEEE Computer Soc. Press, 1995.
Abstract: The recent literature on Markov Regenerative Stochastic Petri Nets (MRSPN) assumes that the random firing time associated to each transition is resampled each time the transition fires or is disabled by the firing of a competitive transition. This modeling assumption does not cover the case of preemption mechanisms of repeat indentical nature (pri). In this policy, an interrupted job must be repeated with an identical requirement so that its associated random variable must not be resampled. The paper investigates the implication of a pri policy into a MRSPN and describes an analytical procedure for the derivation of expressions for the transient probabilities.
Keywords: Stochastic Petri Nets; Semi-Markov Reward Models; Markov regenerative processes; preemptive repeat identical policy.