Decentralized Planning Using Probabilistic Hyperproperties

článek ve sborníku mimo WoS a Scopus

angličtina

Multi-agent planning under stochastic dynamics is usually formalised using decentralized (partially observable) Markov decision processes (MDPs) and reachability or expected reward specifications. In this paper, we propose a different approach: we use an MDP describing how a single agent operates in an environment and probabilistic hyperproperties to capture desired temporal objectives for a set of decentralized agents operating in the environment. We extend existing approaches for model checking probabilistic hyperproperties to handle temporal formulae relating paths of different agents, thus requiring the self-composition between multiple MDPs. Using several case studies, we demonstrate that our approach provides a flexible and expressive framework to broaden the specification capabilities with respect to existing planning techniques. Additionally, we establish a close connection between a subclass of probabilistic hyperproperties and planning for a particular type of Dec-MDPs, for both of which we show undecidability. This lays the ground for the use of existing decentralized planning tools in the field of probabilistic hyperproperty verification.

Probabilistic Hyperproperties, Decentralized Planning, Markov Decision Processes, Abstraction Refinement, Self-composition

ANDRIUSHCHENKO, R.; ČEŠKA, M.; MACÁK, F.; FRANCESCO, P.; MICHELE, C.

26. 5. 2025

Detroit

979-8-4007-1426-9

Proc. of the 24th International Conference on Autonomous Agents and Multiagent Systems

1688

1697

10