Learning Timed Automata via Genetic Programming
Model learning has gained increasing interest in recent years. It derives behavioural models from test data of black-box systems. The main advantage offered by such techniques is that they enable model-based analysis without access to the internals of a system. Applications range from testing to model checking and system understanding. Current work focuses on learning variations of finite state machines. However, most techniques consider discrete time. In this paper, we introduce and demonstrate a method for learning timed automata, finite state machines extended with real-valued clocks. The method is passive, i.e. it generates a model consistent with a set of timed traces produced by a black-box timed system, but does not actively query the system. It is based on genetic programming, a search-based method for automatic program creation. We evaluate our approach on four manually created timed systems and on 40 randomly generated timed automata
READ FULL TEXT