A Two-Player Zero-sum Game Where Only One Player Observes a Brownian Motion
We study a two-player zero-sum game in continuous time, where the payoff—a running cost—depends on a Brownian motion. This Brownian motion is observed in real time by one of the players. The other one observes only the actions of his/her opponent. We prove that the game has a value and characterize it as the largest convex subsolution of a Hamilton–Jacobi equation on the space of probability measures.
KeywordsZero-sum continuous-time game Incomplete information Hamilton–Jacobi equations Brownian motion Measure-valued process
Mathematics Subject Classification91A05 91A23 49N70
We thank Pierre Cardaliaguet for very fruitful discussions. We thank the referees for their careful reading and their pertinent remarks. This work has been partially supported by the French National Research Agency ANR-16-CE40-0015-01 MFG.
- 1.Aumann RJ, Maschler MB (1995) Repeated games with incomplete information. With the collaboration of Richard E. Stearns. MIT Press, CambridgeGoogle Scholar
- 4.Cardaliaguet P Notes on Mean Field Games, unpublishedGoogle Scholar
- 6.Cardaliaguet P, Delarue F, Lasry JF, Lions PL (2015) The master equation and the convergence problem in mean field games. arXiv:1509.02505
- 15.Gensbittel F, Rainer C (2016) A probabilistic representation for the value of zero-sum differential games with incomplete information on both sides. arXiv:1602.06140
- 19.Lions P-L (2005–2006) Cours au Collège de France. http://www.college-de-france.fr
- 24.Sorin S (2016) Limit value of dynamic zero-sum games with vanishing stage duration. arXiv:1603.09089
- 25.Villani C (2003) Topics in optimal transportation, Graduate Studies in Mathematics, 58. American Mathematical Society, ProvidenceGoogle Scholar