scientific article; zbMATH DE number 3861015
[For the entire collection see Zbl 0527.00042.] This work consists in the study of a zero-sum stochastic game whose cost is: \[ E(X_ s1_{(S\leq T)}-X'_ T1_{(T<S)}+(C^ u-C^{\prime v})_{S\wedge T}) \] where the first (resp. second) player chooses (u,S) (resp. (v,T)) with u (resp. v) continuous non anticipative strategy, and S (resp. T) stopping strategy. Under Mokobodski's generalized assumptions, i.e. there exist two positive bounded supermartingale- systems \(\tilde Z^ u\) and \(\tilde Z^{\prime u}\) such that \(C^ u+X\leq Z^ u-Z^{\prime u}\leq -C^{\prime u}-X\), we prove the existence of a value for the game, since \(C^ u,C^{\prime u}\) for all u, X, X' are upper-right semicontinuous and there exists two strategies \(u_ 0\), \(v_ 0\) such that \(C^ u_ 0\) and \(C^{\prime v}_ 0\) are right continuous. Furthermore if for all strategies u: \(C^ u_ t=\int^{t}_{0}e^{-\alpha s}C(s,u_ s)ds\), \(C_ t^{\prime u}=\int^{t}_{0}e^{-\alpha s}C'(s,u_ s)ds\) with C positive bounded, there exists a saddle-point for the game. We use in this paper the results on mixed conrol of \textit{N. El Karoui} [Lect. Notes Math. 721, 443-452 (1979; Zbl 0411.60077)].
- Mixed generalized Dynkin game and stochastic control in a Markovian framework
- scientific article; zbMATH DE number 3897939 (Why is no real title available?)
- Dynkin games and martingale methods
- scientific article; zbMATH DE number 3992548 (Why is no real title available?)
- Continuous-Time Dynkin Games with Mixed Strategies
This page was built for publication:
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3329334)