On a stopping game in continuous time (Q2809215)

The paper deals with a stopping game where both players' stopping decisions have impact on the payoff. On a filtered probability space \((\Omega,\mathcal{F},P,\mathbb{F}=(\mathcal{F}_t)_{t=0,\ldots,T})\), a gain function \(U(s,t)\) is defined, which is \(\mathcal{F}_{s\vee t}\)-measurable. Let \(\mathcal{T}\) be the set of stopping times. \(\rho\) and \(\tau\) are mappings from \(\mathcal T\;\) to \(\mathcal{T}\) satisfying certain non-anticipativity conditions. Define \(\overline C:=\inf_{\rho}\sup_{\tau\in\mathcal T}\mathbf{E}[U(\rho(\tau),\tau)]\) and \(\underline C:=\sup_{\tau}\inf_{\rho\in\mathcal T}\mathbf{E}[U(\rho,\tau(\rho))]\). By converting the problems into a corresponding Dynkin game it is shown that \(\overline C=\underline C\). The results can be applied to weaken the assumption on path regularities on the reward processes in a paper by \textit{M. Kobylanski} et al. [Stochastics 86, No. 2, 304--329 (2014; Zbl 1298.60050); corrigendum ibid. 86, No. 2, 370 (2014; Zbl 1298.60051)].