Solving MDPs with Skew Symmetric Bilinear Utility Functions - Inria - Institut national de recherche en sciences et technologies du numérique Accéder directement au contenu
Communication Dans Un Congrès Année : 2015

Solving MDPs with Skew Symmetric Bilinear Utility Functions

Hugo Gilbert
  • Fonction : Auteur
  • PersonId : 971449
Olivier Spanjaard
Paolo Viappiani

Résumé

In this paper we adopt Skew Symmetric Bilinear (SSB) utility functions to compare policies in Markov Decision Processes (MDPs). By considering pairs of alternatives, SSB utility theory generalizes von Neumann and Morgenstern’s expected utility (EU) theory to encompass rational decision behaviors that EU cannot accommodate. We provide a game-theoretic analysis of the problem of identifying an SSB-optimal policy in finite horizon MDPs and propose an algorithm based on a double oracle approach for computing an optimal (possibly randomized) policy. Finally, we present and discuss experimental results where SSB-optimal policies are computed for a popular TV contest according to several instantiations of SSB utility functions.
Fichier principal
Vignette du fichier
pub_SSB.pdf (330.85 Ko) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)
Loading...

Dates et versions

hal-01212802 , version 1 (30-06-2017)

Identifiants

  • HAL Id : hal-01212802 , version 1

Citer

Hugo Gilbert, Olivier Spanjaard, Paolo Viappiani, Paul Weng. Solving MDPs with Skew Symmetric Bilinear Utility Functions. 24th International Joint Conference on Artificial Intelligence (IJCAI-15), Jul 2015, Buenos Aires, Argentina. pp.1989-1995. ⟨hal-01212802⟩
143 Consultations
95 Téléchargements

Partager

Gmail Facebook X LinkedIn More