November 2021 CDT in Maths of Random Systems Workshop
This event will take place in person in the Huxley Building, room 139.
For a link to view the talks remotely please contact lydia.noa@imperial.ac.uk
3.00-3.40 Benedikt Petko Brownian motion on Riemannian manifolds
3.40-4.15 Felix Prenzel Data-driven modelling of limit order markets
4.15-5.00 Dr Dante Kalise High-dimensional approximation of Hamilton-Jacobi-Bellman PDEs – architectures, algorithms and applications
Hamilton-Jacobi Partial Differential Equations (HJ PDEs) are a central object in optimal control and differential games, enabling the computation of robust controls in feedback form. High-dimensional HJ PDEs naturally arise in the feedback synthesis for high-dimensional control systems, and their numerical solution must be sought outside the framework provided by standard grid-based discretizations. In this talk, I will discuss the construction novel computational methods for approximating high-dimensional HJ PDEs, based on tensor decompositions, polynomial approximation, and deep neural networks.
CDT Student (Imperial College London)
Benedikt is a PhD student on our CDT in mathematics at Imperial College London. He studies an intersection of probability theory, Riemannian and discrete geometry. He works under the supervision of Prof Xue-Mei Li who is a specialist in stochastic differential geometry.
CDT Student (University of Oxford)
Felix is a DPhil student on the CDT working at the University of Oxford Mathematical Institute. He works with industry partners JP Morgan with supervision from Prof Mihai Cucuringu.
Imperial College London
Dante Kalise is Senior Lecturer in Computational Optimisation and Control at the Department of Mathematics, Imperial College London. Before joining Imperial, he was Assistant Professor in Applied Mathematics at the University of Nottingham, and held research positions at RICAM Linz, and at La Sapienza University of Rome. Kalise's research spans topics in scientific computing, optimal control, and PDEs. His current research is focused on the analysis and design of computational methods for the solution of high-dimensional HJB PDEs and applications in nonlinear feedback control for PDE dynamics and agent-based models.
