Computational effort of BDD-based supervisor synthesis of extended finite automata

We consider supervisor synthesis of Extended Finite Automata that are represented using Binary Decision Diagrams (BDDs). Peak used BDD nodes and BDD operation count are introduced as platform independent and deterministic metrics that quantitatively indicate the computational effort needed to synthesize a supervisor. The use of BDD operation count is novel with respect to expressing supervisor synthesis effort. The (dis-)advantages of using these metrics to state of practice metrics such as wall clock time and worst case state space size are analyzed. The supervisor synthesis algorithm is initiated with a certain event- and variable order. It is already known from literature that variable order influences synthesis performance. We show that the event order is also relevant to consider. We discuss how these orders influence the synthesis effort and, by performing an experiment on a set of models, we show the extent of this influence.