David Criens
University of Freiburg
Room: M3 3127
On dynamic risk measures and variational formulas
The celebrated Boue-Dupuis formula provides a variational optimal control representation for entropic transformations of Brownian functionals. A fundamental application is an elegant approach to Laplace principles for the small-noise and risk-sensitivity analysis of diffusion models. The classical proof for the Boue-Dupuis formula is based on dual representations, Girsanov-type arguments for probabilities that are dominated by the Wiener measure, and approximation techniques. In this talk we explain a new approach to derive the Boue-Dupuis formula from intrinsic properties of dynamic risk measures and standard viscosity theory for Hamilton-Jacobi-Bellman equations.
The talk is based on joint work with Michael Kupper (University of Konstanz):
Representation Theorems for Convex Expectations and Semigroups on Path Space on arXiv.