Optimal Stopping with Dynamic Variational Preferences

We consider optimal stopping problems in uncertain environments for an agent assessing utility by virtue of dynamic variational preferences or, equivalently, assessing risk by dynamic convex risk measures. The solution is achieved by generalizing the approach in terms of multiple priors introducing the concept of variational supermartingales and an accompanying theory. To illustrate results, we consider prominent examples: dynamic entropic risk measures and a dynamic version of generalized average value at risk.