Abstract

We present an active automata learning algorithm for Mealy state machines that directly constructs a state machine hypothesis according to observations, while other algorithms generate a state machine as output from information gathered in an observation table. Our DHC algorithm starts with a one-state hypothesis that it successively extends using a direct construction approach. This approach enables direct observation of the automata construction process: the learning algorithm continues to complete its hypothesis, providing intuition to a field of formal methods otherwise dominated by algorithms that largely operate on internal data structures without visible feedback.

The DHC algorithm is competitive in cases where memory is the critical issue, e.g., in embedded networked systems. It is also well-suited as educational tool to teach the underlying well-established theoretical methods in a totally unbiased fashion, without cluttering the view onto the actual idea of the learning process with aspects only relevant to internal bookkeeping.