Learning a Partially-Known Discrete Event System

There are many cases in which our understanding of a system may be limited due to its complexity or lack of access into the entire system, leaving us with only partial system knowledge. This paper proposes a novel systematic active-learning method for realizing a partially-known Discrete Event System (DES). The proposed technique takes the available information about the system into account by tabularly capturing the known data from the system, and then, discovers the unknown part of the system via an active-learning procedure. For this purpose, a series of tables will be constructed to first infer the information about the system from the available data, and if unavailable, the developed algorithm collects the information through basic queries made to an oracle. It is proven that the developed technique returns a language-equivalent finite-state automaton model for the system under identification after a finite number of iterations. A real-world illustrative example is provided to explain the details of the proposed method.