Partial-Order Reduction for Supervisory Controller Synthesis

A key challenge in the synthesis and subsequent analysis of supervisory controllers is impact state-space explosion caused by concurrency. The main bottleneck often memory needed to store composition plant requirement automata resulting supervisor. Partial-order reduction (POR) a well-established technique that alleviates this issue field model checking. It does so exploiting redundancy with respect properties interest. For controller synthesis, functional interest are nonblockingness, controllability, least-restrictiveness, but also performance properties, such as throughput latency We propose an on-the-fly POR on input preserves both synthesized controller. This improves scalability (and any analysis). Synthesis experiments show effectiveness set realistic manufacturing system models.