Performance evaluation using measure dependent transitions in Petri nets

Decomposition is an often proposed method to determine Petri nets with a large state space. But there are no generalized rules on how to decompose a Petri net. We present a general method to adapt the transition parameters of timed Petri nets that show a cyclic behavior. The cyclic time intervals can be of constant length or differ. We realize this by adapting the firing rates of exponentially timed transitions or the weights of immediate transitions. Our method is demonstrated with two applications. They show the acceleration in the calculation time. But also, the sensitivity to neglected dependences in the decomposed Petri net can be observed.