Reinforcement learning and differentiable simulations for autonomous tuning and control of linear particle accelerators

Publikationstyp
Doctoral Thesis
Date Issued
2026
Sprache
English
Author(s)
Kaiser, Jan  
Advisor
Eichler, Annika  
Referee
Fey, Görschwin  orcid-logo
Title Granting Institution
Technische Universität Hamburg
Place of Title Granting Institution
Hamburg
Examination Date
2025-06-04
Institute
Regelungstechnik E-14  
TORE-DOI
10.15480/882.16060
TORE-URI
https://hdl.handle.net/11420/58370
Citation
Dr. Hut 978-3-8439-5683-3: (2025)
Publisher
Dr. Hut
ISBN
978-3-8439-5683-3
Particle accelerators are sophisticated scientific facilities that require precise but time-consuming optimisation to achieve optimal performance. Considering benchmark tasks at the ARES and LCLS facilities, this dissertation proposes methods to deploy simulation-trained reinforcement learning (RL) policies for accelerator tuning zero-shot to the real world and novel tuning tasks, while comparing their performance to traditional methods. A high-speed differentiable beam dynamics simulator is developed to make collecting large datasets for RL feasible, and to enable a multitude of novel gradient-based accelerator applications. These contributions lay the groundwork for faster accelerator tuning to better working points, and enable new scientific discoveries.
Subjects
reinforcement learning
differentiable simulation
particle accelerators
DDC Class
539: Matter; Molecular Physics; Atomic and Nuclear physics; Radiation; Quantum Physics
006.31: Machine Learning
681.2: Testing, Measuring, Sensing Instruments
Lizenz
https://creativecommons.org/licenses/by/4.0/
Loading...
Thumbnail Image
Name

Kaiser_Jan_Reinforcement-Learning-and-Differentiable-Simulations-for-Autonomous-Tuning-and-Control-of-Linear-Particle-Accelerators.pdf

Size

42.4 MB

Format

Adobe PDF

Version History

Now showing1 - 1 of 1
VersionDateSummary
1*
 2026-02-12 10:03:55
* Selected version