raxml: benchmarks for automata learning
This repository contains benchmarks for automata learning algorithms. The benchmarks are register automaton models and come from a variety of sources and application domains. They range from explanatory examples that have been discussed in the literature, to manually written specifications of data structures, to models that were inferred from actual systems (e.g., the Biometric Passport and the SIP protocol). For further documentation on the benchmarks please checkout the papers in the ./papers folder.
The benchmarks distributed under the terms of the BSD license (see license.txt).
The Java parser for the benchmarks is distributed under LGPLv3 (see http://www.gnu.de/documents/lgpl.en.html).
How To Use
This repository contains a set of benchmark register automata in an xml format. The corresponding schema is provided as well. A description of the intended semantics of the automaton format will hopefully be added in the future. Until then, the repository provides a Java implementation that can parse the provided benchmarks and execute data words on these automata.
- The ./benchmarks folder contains benchmarks, scripts that create xml files for data structures of varying size, and the xsd schema.
- The folder ./rajava contains a Java parser for the benchmarks.
- The folder ./papers contains additonal documentation.
How To Use
There is some documentation available at http://www.mbsd.cs.ru.nl/automata for the different benchmarks.