Learning Behavior Models for Hybrid Timed Systems

A tailored model of a system is the prerequisite for various analysis tasks, such as anomaly detection, fault identification, or quality assurance. This paper deals with algorithmic learning system’s behavior given sample observations. In particular, we consider real-world production plants where learned must capture timing behavior, dependencies between variables, well mode switches—in short: hybrid characteristics. Usually, formation tasks are solved by human engineers, entailing well-known bunch problems including knowledge acquisition, development cost, lack experience. Our contributions to outlined field follows. (1) We present taxonomy related tasks. As result, an important open problem domain identified: The timed automata. (2) For this class models, algorithm HyBUTLA presented. first its kind solve underlying at scalable precision. (3) two case studies that illustrate usability approach in realistic settings. (4) give proof and runtime properties HyBUTLA.