Dresden 2026 – scientific programme

Parts | Days | Selection | Search | Updates | Downloads | Help

AKPIK: Arbeitskreis Physik, moderne Informationstechnologie und Künstliche Intelligenz

AKPIK 5: Poster

AKPIK 5.4: Poster

Thursday, March 12, 2026, 15:00–16:30, P5

Exploring Reinforcement Learning for Particle Transport in the Presence of Inhomogeneities — •Finn Marten Boyer, Atreya Majumdar, and Karin Everschor-Sitte — University Duisburg-Essen, Duisburg, Germany

Classical transport theories typically assume homogeneous media, yet real materials often exhibit inhomogeneities that limit the applicability of such models. In particular, standard approaches like renormalization may fail when particles encounter defects whose characteristic energy scales are comparable to or larger than their kinetic energies. We investigate reinforcement learning as a data-driven framework for optimizing particle transport in strongly inhomogeneous environments. Our work indicates the potential of reinforcement-learning-based approaches for particle dynamics in more realistic and complex systems.

Keywords: Transport; Reinforcement Learning; Particles; Data-driven modeling