2015
On Frequency LTL in Probabilistic Systems
FOREJT, Vojtěch a Jan KRČÁLZákladní údaje
Originální název
On Frequency LTL in Probabilistic Systems
Autoři
FOREJT, Vojtěch (203 Česká republika, domácí) a Jan KRČÁL (203 Česká republika)
Vydání
Madrid, Spain, CONCUR 2015, od s. 184-197, 14 s. 2015
Nakladatel
Schloss Dagstuhl - Leibniz-Zentrum fuer Informatik
Další údaje
Jazyk
angličtina
Typ výsledku
Stať ve sborníku
Obor
10201 Computer sciences, information science, bioinformatics
Stát vydavatele
Německo
Utajení
není předmětem státního či obchodního tajemství
Forma vydání
elektronická verze "online"
Kód RIV
RIV/00216224:14330/15:00081289
Organizační jednotka
Fakulta informatiky
ISBN
978-3-939897-91-0
ISSN
Klíčová slova anglicky
markov chains; markov decision processes; ltl; controller synthesis
Štítky
Příznaky
Mezinárodní význam, Recenzováno
Změněno: 28. 4. 2016 15:33, RNDr. Pavel Šmerk, Ph.D.
Anotace
V originále
We study frequency linear-time temporal logic (fLTL) which extends the linear-time temporal logic (LTL) with a path operator G^p expressing that on a path, certain formula holds with at least a given frequency p, thus relaxing the semantics of the usual G operator of LTL. Such logic is particularly useful in probabilistic systems, where some undesirable events such as random failures may occur and are acceptable if they are rare enough. Frequency-related extensions of LTL have been previously studied by several authors, where mostly the logic is equipped with an extended "until" and "globally" operator, leading to undecidability of most interesting problems. For the variant we study, we are able to establish fundamental decidability results. We show that for Markov chains, the problem of computing the probability with which a given fLTL formula holds has the same complexity as the analogous problem for LTL. We also show that for Markov decision processes the problem becomes more delicate, but when restricting the frequency bound p to be 1 and negations not to be outside any G^p operator, we can compute the maximum probability of satisfying the fLTL formula. This can be again performed with the same time complexity as for the ordinary LTL formulas.
Návaznosti
GBP202/12/G061, projekt VaV |
|