On generation of state space for timed Petri nets

Abstract

It is shown that the behavior of timed Petri nets with deterministic firing times (D-timed nets) and with exponentially distributed random firing times (M-timed nets) can be described within one uniform formalism. Moreover, for both classes of nets the state spaces are homogeneous semi-Markov chains, the stationary probabilities of states and many performance measures can thus be obtained by standard techniques developed for analysis of Markov processes. Because of scarcity of nets as well as corresponding systems of equilibrium equations, list structure representations are proposed, and a general procedure for generation of the state space is outlined to show the required processing of list structures.