One-Counter Markov Decision Processes

TL;DR

This paper investigates the computational complexity of analysis problems for One-Counter Markov Decision Processes, providing new bounds and algorithms for key questions like termination and limit behaviors.

Contribution

It introduces complexity bounds and algorithms for analysis problems in OC-MDPs, connecting them to queueing theory, automata, and stochastic models.

Findings

Polynomial-time algorithms for some problems.

PSPACE- and BH-hardness results for others.

EXPTIME upper bounds established.

Abstract

We study the computational complexity of central analysis problems for One-Counter Markov Decision Processes (OC-MDPs), a class of finitely-presented, countable-state MDPs. OC-MDPs are equivalent to a controlled extension of (discrete-time) Quasi-Birth-Death processes (QBDs), a stochastic model studied heavily in queueing theory and applied probability. They can thus be viewed as a natural ``adversarial'' version of a classic stochastic model. Alternatively, they can also be viewed as a natural probabilistic/controlled extension of classic one-counter automata. OC-MDPs also subsume (as a very restricted special case) a recently studied MDP model called ``solvency games'' that model a risk-averse gambling scenario. Basic computational questions about these models include ``termination'' questions and ``limit'' questions, such as the following: does the controller have a ``strategy'' (or ``policy'') to ensure that the counter (which may for example count the number of jobs in the queue) will hit value 0 (the empty queue) almost surely (a.s.)? Or that it will have infinite limsup value, a.s.? Or, that it will hit value 0 in selected terminal states, a.s.? Or, in case these are not satisfied a.s., compute the maximum (supremum) such probability over all strategies. We provide new upper and lower bounds on the complexity of such problems. For some of them we present a polynomial-time algorithm, whereas for others we show PSPACE- or BH-hardness and give an EXPTIME upper bound. Our upper bounds combine techniques from the theory of MDP reward models, the theory of random walks, and a variety of automata-theoretic methods.