A symbolic approach to large-scale discrete event systems modeled as finite automata with variables

The state-space explosion problem, resulting from the reachability computation of the synthesis task, is one of the main obstacles preventing the supervisory control theory (SCT) from having an industrial breakthrough. To alleviate this problem, a well-known strategy is to utilize binary decision diagrams (BDDs) to compute supervisors symbolically. Based on this principle, we present in this paper an efficient reachability approach to large-scale discrete event systems modeled as finite automata with variables. By making use of the disjunctive partitioning technique, the proposed approach partitions the transition relation of a considered system into a set of partial transition relations according to included events. Then those partial transition relations are selected systematically to perform the reachability computation. Experimental results show that more iterations might be required to compute the fixed point, but the intermediate BDDs are smaller. The approach has been implemented in the supervisory control tool Supremica and the efficiency is demonstrated on a set of industrially relevant benchmark problems.