Computing measures of weak-MSO definable sets of trees

Abstract

This work addresses the problem of computing measures of recognisable sets of infinite trees. An algorithm is provided to compute the probability measure of a tree language recognisable by a weak alternating automaton, or equivalently definable in weak monadic second-order logic. The measure is the uniform coin-flipping measure or more generally it is generated by a~branching stochastic process. The class of tree languages in consideration, although smaller than all regular tree languages, comprises in particular the languages definable in the alternation-free mu-calculus or in temporal logic CTL. Thus, the new algorithm may enhance the toolbox of probabilistic model checking.

An ArXiv version of a paper from ICALP 2020