Decisive Markov chains

Abstract

We consider qualitative and quantitative verification problems for infinitestate Markov chains. We call a Markov chain decisive w.r.t. a given set of target states F if it almost certainly eventually reaches either F or a state from which F can no longer be reached. While all finite Markov chains are trivially decisive (for every set F), this also holds for many classes of infinite Markov chains. Infinite Markov chains which contain a finite attractor are decisive w.r.t. every set F. In particular, all Markov chains induced by probabilistic lossy channel systems (PLCS) contain a finite attractor and are thus decisive. Furthermore, all globally coarse Markov chains are decisive. The class of globally coarse Markov chains includes, e.g., those induced by probabilistic vector addition systems (PVASS) with upward-closed sets F, and all Markov chains induced by probabilistic noisy Turing machines (PNTM) (a generalization of the noisy Turing machines (NTM) of Asarin and Collins). We consider both safety and liveness problems for decisive Markov chains. Safety: What is the probability that a given set of states F is eventually reached. Liveness: What is the probability that a given set of states is reached infinitely often. There are three variants of these questions. (1) The qualitative problem, i.e., deciding if the probability is one (or zero); (2) the approximate quantitative problem, i.e., computing the probability up-to arbitrary precision; (3) the exact quantitative problem, i.e., computing probabilities exactly. 1. We express the qualitative problem in abstract terms for decisive Markov chains, and show an almost complete picture of its decidability for PLCS, PVASS and PNTM. 2. We also show that the path enumeration algorithm of Iyer and Narasimha terminates for decisive Markov chains and can thus be used to solve the approximate quantitative safety problem. A modified variant of this algorithm can be used to solve the approximate quantitative liveness problem. 3. Finally, we show that the exact probability of (repeatedly) reaching F cannot be effectively expressed (in a uniform way) in Tarski-algebra for either PLCS, PVASS or (P)NTM (unlike for probabilistic pushdown automata).