Thursday, May 22, 2025 - 12:00 in V2-210
An efficient numerical approximation of a linear-quadratic stochastic boundary optimal control problem
A talk in the SPDEvent series by
Abhishek Chaudhary from Universität Tübingen
| Abstract: |
This talk presents a fast and implementable discretisation for the Dirichlet boundary control problem associated with the stochastic heat equation and demonstrates its space-time convergence with rates. After space-time discretisation, the discrete optimality conditions involve the discretisation of a backward SPDE, whose numerical solution is typically costly due to the computation of conditional expectations. We propose a reformulation of the discrete optimality conditions that
eliminates the need for simulating conditional expectations, thereby significantly reducing computational complexity compared to
regression-based simulations while maintaining the same convergence rate.
This is joint work with Fabian Merle, Andreas Prohl, and Yanqing Wang.
|
Back
