作者: Taolue Chen , Marco Diciolla , Marta Kwiatkowska , Alexandru Mereacre
DOI: 10.1007/978-3-642-24310-3_4
关键词: Timed automaton 、 Discrete mathematics 、 Pointwise 、 Bounded function 、 Approximation algorithm 、 Metric (mathematics) 、 Interval (mathematics) 、 Markov chain 、 Mathematics 、 Path (graph theory)
摘要: In this paper we study time-bounded verification of a finite continuous-time Markov chain (CTMC) C against real-time specification, provided either as metric temporal logic (MTL) property φ, or timed automaton (TA) A. The key question is: what is the probability set paths that satisfy φ (or are accepted by A) over time interval fixed, bounded length? We provide approximation algorithms to solve these problems. first derive bound N such with at most discrete jumps sufficient approximate desired up e. Then, for each (untimed) path σ length N, generate constraints variables determining residence state along σ, depending on realtime specification under consideration. paths, determined and associated constraints, can thus be formulated multidimensional integral. Summing all probabilities yields result. For MTL, consider both continuous pointwise semantics. differ mainly in generation two types specifications.