Nicolas Markey

Abstract

A weighted timed game is a timed game with extra quantitative information representing e.g. energy consumption. Optimizing the cost for reaching a target is a natural question, which has been investigated for ten years. Existence of optimal strategies is known to be undecidable in general, and only very restricted classes of games have been described for which optimal cost and almost-optimal strategies can be computed.

In this paper, we show that the value problem is undecidable in general weighted timed games. The undecidability proof relies on that for the existence of optimal strategies and on a diagonalization construction recently designed in the context of quantitative temporal logics. We then provide an algorithm to compute arbitrary approximations of the value in a game, and almost-optimal strategies. The algorithm applies in a large subclass of weighted timed games, and is the first approximation scheme which is designed in the current context.

@inproceedings{concur2015-BJM,
  author =              {Bouyer, Patricia and Jaziri, Samy and Markey,
                         Nicolas},
  title =               {On~the Value Problem in Weighted Timed Games},
  editor =              {Aceto, Luca and de Frutos{-}Escrig, David},
  booktitle =           {{P}roceedings of the 26th {I}nternational
                         {C}onference on {C}oncurrency {T}heory
                         ({CONCUR}'15)},
  acronym =             {{CONCUR}'15},
  publisher =           {Leibniz-Zentrum f{\"u}r Informatik},
  series =              {Leibniz International Proceedings in Informatics},
  volume =              {42},
  pages =               {311-324},
  year =                {2015},
  month =               sep,
  doi =                 {10.4230/LIPIcs.CONCUR.2015.311},
  abstract =            {A~weighted timed game is a timed game with extra
                         quantitative information representing e.g. energy
                         consumption. Optimizing the cost for reaching a
                         target is a natural question, which has been
                         investigated for ten years. Existence of optimal
                         strategies is known to be undecidable in general,
                         and only very restricted classes of games have been
                         described for which optimal cost and almost-optimal
                         strategies can be computed.\par In this paper, we
                         show that the value problem is undecidable in
                         general weighted timed games. The undecidability
                         proof relies on that for the existence of optimal
                         strategies and on a diagonalization construction
                         recently designed in the context of quantitative
                         temporal logics. We then provide an algorithm to
                         compute arbitrary approximations of the value in a
                         game, and almost-optimal strategies. The algorithm
                         applies in a large subclass of weighted timed games,
                         and is the first approximation scheme which is
                         designed in the current context.},
}