Learning Nondeterministic Real-Time Automata

We present an active learning algorithm named NRTALearning for nondeterministic real-time automata (NRTAs). Real-time (RTAs) are a subclass of timed with only one clock which resets at each transition. First, we prove the corresponding Myhill-Nerode theorem languages. Then show that there exists unique minimal deterministic automaton (DRTA) recognizing given language, but same does not hold NRTAs. thus define special kind NRTAs, residual (RRTAs), and RRTA to recognize any language. This transforms problem NRTAs RRTAs. After describing in detail, its correctness polynomial complexity. In addition, based on theorem, extend existing NL* finite learn evaluate compare two algorithms benchmarks consisting randomly generated rational regular expressions. The results generally performs fewer membership queries more equivalence than extended algorithm, learnt have much locations DRTAs. also conduct case study using model scheduling final testing integrated circuits.