Publications 2025
Research papers, software, lecture notes

Prev
Next
No. 2025.04
Kinetic variable-sample methods for stochastic optimization problems
S. Bonandin and M. Herty
Subject: Global optimization, stochastic optimization problems, particle-based methods, consensus-based optimization, Boltzmann equation, kinetic equations

Abstract

We discuss variable-sample strategies and consensus- and kinetic-based particle optimization methods for problems where the cost function represents the expected value of a random mapping. Variable-sample strategies replace the expected value by an approximation at each iteration of the optimization algorithm. We introduce a novel variable-sample inspired time-discrete consensus-type algorithm and demonstrate its computational efficiency. Subsequently, we present an alternative time-continuous kinetic-based description of the algorithm, which allows us to exploit tools of kinetic theory to conduct a comprehensive theoretical analysis. Finally, we test the consistency of the proposed modelling approaches through several numerical experiments.

Download

arXiv:2502.17982

No. 2025.03
High-dimensional stochastic finite volumes using the tensor train format
J. Dubois, M. Herty, and S. Müller
Subject: Numerical analysis

Abstract

We propose a method for the uncertainty quantification of nonlinear hyperbolic equations with many uncertain parameters. The method combines the stochastic finite volume method and tensor trains in a novel way: the physical space and time dimensions are kept as full tensors, while all stochastic dimensions are compressed together into a tensor train. The resulting hybrid format has one tensor train for each spatial cell and each time step. We adapt a MUSCL scheme to this hybrid format and show the feasibility of the approach using several classical test cases. A convergence study is done on the Burgers' equation with three stochastic parameters. We also solve the Burgers' equation for an increasing number of stochastic dimensions and show an example with the Euler equations. The presented method opens new avenues for combining uncertainty quantification with well-known numerical schemes for conservation law.

Download

arXiv:2502.04868

No. 2025.02
The Lax-Friedrichs method in one-dimensional hemodynamics
A. Beckers and N. Kolbe
Subject: Blood flow modeling, cardiovascular networks, finite volumes, coupled conservation laws, boundary conditions, hyperbolic systems

Abstract

The discretization of reduced one-dimensional hyperbolic models of blood flow using the Lax-Friedrichs method is discussed. Employing the well-established central scheme in this domain significantly simplifies the implementation of specific boundary and coupling conditions in vascular networks accounting e.g. for a periodic heart beat, vascular occlusions, stented vessel segments and bifurcations. In particular, the coupling of system extensions modeling patient specific geometries and therapies can be realized without information on the eigenstructure of the models. For the derivation of the scheme and the coupling conditions a relaxation of the model is considered and its discrete relaxation limit evaluated. Moreover, a second order MUSCL-type extensions of the scheme is introduced. Numerical experiments in uncoupled and coupled cases that verify the consistency and convergence of the approach are presented.

Download

arXiv:2501.16115

No. 2025.01
On the relation between approaches for boundary feedback control of hyperbolic systems
M. Herty and F. Thein
Subject: Boundary feedback control, stabilization, hyperbolic systems of PDEs, Lyapunov analysis, linear matrix inequality

Abstract

Stabilization of partial differential equations is a topic of utmost importance in mathematics as well as in engineering sciences. Concerning one dimensional problems there exists a well developed theory. Due to numerous important applications the interest in boundary feedback control of multi-dimensional hyperbolic systems is increasing. In the present work we want to discuss the relation between some of the most recent results available in the literature. The key result of the present work is to show that the type of system discussed in Yang and Yong (2024) identifies a particular class which falls into the framework presented in Herty and Thein (2024).

Reference

Eur. J. Control 82 (2025), Paper No. 101182

Prev
Next