Highlights of the Year
New Software and Platforms
Bilateral Contracts and Grants with Industry
Partnerships and Cooperations
Bibliography
 PDF e-Pub

Section: New Results

New results on Axis 1: Quantitative models

Verification of Real-Time Models

Participants : Ocan Sankur, Nicolas Markey, Victor Roussanaly

Abstraction-refinement algorithms for model checking of timed automata.

The abstraction domain we consider [26] abstracts away zones by restricting the set of clock constraints that can be used to define them, while the refinement procedure computes the set of constraints that must be taken into consideration in the abstraction so as to exclude a given spurious counterexample. We implement this idea in two ways: an enumerative algorithm where a lazy abstraction approach is adopted, meaning that possibly different abstract domains are assigned to each exploration node; and a symbolic algorithm where the abstract transition system is encoded with Boolean formulas.

Robust controller synthesis problem in Büchi timed automata

We solve a robust controller synthesis problem [20] in a purely symbolic way. The goal of the controller is to play according to an accepting lasso of the automaton, while resisting to timing perturbations chosen by a competing environment. The problem was previously shown to be PSPACE-complete using regions-based techniques, but we provide a first tool solving the problem using zones only, thus more resilient to state-space explosion problem. The key ingredient is the introduction of branching constraint graphs allowing to decide in polynomial time whether a given lasso is robust, and even compute the largest admissible perturbation if it is. We also make an original use of constraint graphs in this context in order to test the inclusion of timed reachability relations, crucial for the termination criterion of our algorithm. Our techniques are illustrated using a case study on the regulation of a train network.

Verification of Stochastic Models

Participants : Hugo Bazille, Nathalie Bertrand, Éric Fabre, Blaise Genest, Ocan Sankur

Long-run satisfaction of path properties

We introduced the concepts of long-run frequency of path properties for paths in Kripke structures, and their generalization to long-run probabilities for schedulers in Markov decision processes [13]. We then studied the natural optimization problem of computing the optimal values of these measures, when ranging over all paths or all schedulers, and the corresponding decision problem when given a threshold. The main results are as follows. For (repeated) reachability and other simple properties, optimal long-run probabilities and corresponding optimal memoryless schedulers are computable in polynomial time. When it comes to constrained reachability properties, memoryless schedulers are no longer sufficient, even in the non-probabilistic setting. Nevertheless, optimal long-run probabilities for constrained reachability are computable in pseudo-polynomial time in the probabilistic setting and in polynomial time for Kripke structures. Finally for co-safety properties expressed by NFA, we gave an exponential-time algorithm to compute the optimal long-run frequency, and proved the PSPACE-completeness of the threshold problem.

Approximate Verification of Dynamic Bayesian Networks.

We are interested in studying the evolution of large homogeneous populations of cells, where each cell is assumed to be composed of a group of biological players (species) whose dynamics is governed by a complex biological pathway, identical for all cells. Modeling the inherent variability of the species concentrations in different cells is crucial to understand the dynamics of the population. In [9], we focus on handling this variability by modeling each species by a random variable that evolves over time. This appealing approach runs into the curse of dimensionality since exactly representing a joint probability distribution involving a large set of random variables quickly becomes intractable as the number of variables grows. To make this approach amenable to biopathways, we explore different techniques to (i) approximate the exact joint distribution at a given time point, and (ii) to track its evolution as time elapses.

Classification among stochastic systems

An important task in AI is one of classifying an observation as belonging to one class among several (e.g. image classification). We revisit this problem in a verification context: given $k$ partially observable systems modeled as Hidden Markov Models (HMMs, also called labeled Markov chains), and an execution of one of them, can we eventually classify which system performed this execution, just by looking at its observations? Interestingly, this problem generalizes several problems in verification and control, such as fault diagnosis and opacity. Also, classification has strong connections with different notions of distances between stochastic models.

In [12], we study a general and practical notion of classifiers, namely limit-sure classifiers, which allow misclassification, i.e. errors in classification, as long as the probability of misclassification tends to 0 as the length of the observation grows. To study the complexity of several notions of classification, we develop techniques based on a simple but powerful notion of stationary distributions for HMMs. We prove that one cannot classify among HMMs iff there is a finite separating word from their stationary distributions. This provides a direct proof that classifiability can be checked in PTIME, as an alternative to existing proofs using separating events (i.e. sets of infinite separating words) for the total variation distance. Our approach also allows us to introduce and tackle new notions of classifiability which are applicable in a security context.

Fault diagnosis for stochastic systems

Diagnosis of partially observable stochastic systems prone to faults was introduced in the late nineties. Diagnosability, i.e. the existence of a diagnoser, may be specified in different ways: exact diagnosability requires that almost surely a fault is detected and that no fault is erroneously claimed; approximate diagnosability tolerates a small error probability when claiming a fault; last, accurate approximate diagnosability guarantees that the error probability can be chosen arbitrarily small.

In the article [7], we first refine the specification of diagnosability by identifying three criteria: (1) detecting faulty runs or providing information for all runs (2) considering finite or infinite runs, and (3) requiring or not a uniform detection delay. We then give a complete picture of relations between the different diagnosability specifications for probabilistic systems and establish characterisations for most of them in the finite-state case. Based on these characterisations, we develop decision procedures, study their complexity and prove their optimality. We also design synthesis algorithms to construct diagnosers and we analyse their memory requirements. Finally we establish undecidability of the diagnosability problems for which we provided no characterisation.

Energy Games

Participants : Loïc Hélouët, Nicolas Markey

Games with reachability objectives under energy constraints.

Under strict energy constraints (either only lower-bound constraint or interval constraint), we prove [23] that games with reachability objectives are LOGSPACE-equivalent to energy games with the same energy constraints but without reachability objective (i.e., for infinite runs). We then consider two kinds of relaxations of the upper-bound constraints (while keeping the lower-bound constraint strict): in the first one, called weak upper bound, the upper bound is absorbing, in the sense that it allows receiving more energy when the upper bound is already reached, but the extra energy will not be stored; in the second one, we allow for temporary violations of the upper bound, imposing limits on the number or on the amount of violations. We prove that when considering weak upper bound, reachability objectives require memory, but can still be solved in polynomial-time for one-player arenas; we prove that they are in co-NP in the two-player setting. Allowing for bounded violations makes the problem PSPACE-complete for one-player arenas and EXPTIME-complete for two players.