Learning Stochastic Timed Automata from Sample Executions

• Published in: Leveraging Applications of Formal Methods, Verification and Validation. Technologies for Mastering Change pp. 508 - 523
• Main Authors: de Matos Pedro, André, Crocker, Paul Andrew, de Sousa, Simão Melo
• Format: Book Chapter
• Language: English
• Published: Berlin, Heidelberg Springer Berlin Heidelberg 2012
• Series: Lecture Notes in Computer Science
• Subjects: Discrete Event System; Land Mobile Satellite; Propose Learning Algorithm; Sample Clock; Statistical Model Check
• ISBN: 9783642340253; 3642340253
• ISSN: 0302-9743; 1611-3349
• DOI: 10.1007/978-3-642-34026-0_38

Generalized semi-Markov processes are an important class of stochastic systems which are generated by stochastic timed automata. In this paper we present a novel methodology to learn this type of stochastic timed automata from sample executions of a stochastic discrete event system. Apart from its theoretical interest for machine learning area, our algorithm can be used for quantitative analysis and verification in the context of model checking. We demonstrate that the proposed learning algorithm, in the limit, correctly identifies the generalized semi-Markov process given a structurally complete sample. This paper also presents a Matlab toolbox for our algorithm and a case study of the analysis for a multi-processor system scheduler with uncertainty in task duration.