Bisimulation for Markov Decision Processes through Families of Functional Expressions - Inria - Institut national de recherche en sciences et technologies du numérique Accéder directement au contenu
Communication Dans Un Congrès Année : 2014

Bisimulation for Markov Decision Processes through Families of Functional Expressions

Résumé

We transfer a notion of quantitative bisimilarity for labelled Markov processes to Markov decision processes with continuous state spaces. This notion takes the form of a pseudometric on the system states, cast in terms of the equivalence of a family of functional expressions evaluated on those states and interpreted as a real-valued modal logic. Our proof amounts to a slight modification of previous techniques used to prove equivalence with a fixed-point pseudometric on the state-space of a labelled Markov process and making heavy use of the Kantorovich probability metric. Indeed, we again demonstrate equivalence with a fixed-point pseudometric defined on Markov decision processes; what is novel is that we recast this proof in terms of integral probability metrics [5] defined through the family of functional expressions, shifting emphasis back to properties of such families. The hope is that a judicious choice of family might lead to something more computationally tractable than bisimilarity whilst maintaining its pleasing theoretical guarantees. Moreover, we use a trick from descriptive set theory to extend our results to MDPs with bounded measurable reward functions, dropping a previous continuity constraint on rewards and Markov kernels.
Fichier non déposé

Dates et versions

hal-01098566 , version 1 (26-12-2014)

Identifiants

Citer

Norman Ferns, Sophia Knight, Doina Precup. Bisimulation for Markov Decision Processes through Families of Functional Expressions. Horizons of the Mind. A Tribute to Prakash Panangaden (for his 60th birthday), Franck van Breugel; Elham Kashefi; Castucia Palamidessi; Jan Rutten, May 2014, Oxford, United Kingdom. pp.319-342, ⟨10.1007/978-3-319-06880-0_17⟩. ⟨hal-01098566⟩
306 Consultations
0 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More