Markov automata are a powerful formalism for modelling systems which exhibit nondeterminism, probabilistic choices and continuous stochastic timing. We consider the computation of long-run average rewards, the most classical problem in continuous-time Markov model analysis. We propose an algorithm based on value iteration. It improves the state of the art by orders of magnitude. The contribution is rooted in a fresh look on Markov automata, namely by treating them as an efficient encoding of CTMDPs with – in the worst case – exponentially more transitions.
CITATION STYLE
Butkova, Y., Wimmer, R., & Hermanns, H. (2017). Long-run rewards for markov automata. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 10206 LNCS, pp. 188–203). Springer Verlag. https://doi.org/10.1007/978-3-662-54580-5_11
Mendeley helps you to discover research relevant for your work.