Solving finite time horizon Dynkin games by optimal switching

This paper studies the connection between Dynkin games and optimal switching in continuous time and on a finite horizon. An auxiliary two-mode optimal switching problem is formulated which enables the derivation of the game's value under very mild assumptions. Under slightly stronger assumptions, the optimal switching formulation is used to prove the existence of a saddle point and a connection is made to the classical "Mokobodski's hypothesis". Results are illustrated by comparison to numerical solutions of three specific Dynkin games which have appeared in recent papers, including an example of a game option with payoff dependent on a jump-diffusion process.