HZB is currently designing the lattice for BESSY III, the successor of the 1.7 GeV electron storage ring running in Berlin since 1998. HZB follows a deterministic lattice design strategy, where the natural substructures of a non-hybrid MBA lattice are optimized separately. The substructures consist of only a few parameters, that can be derived from the strategic goals of the project. In the next step, the focusing and de-focusing sextupole families are split up, to optimize the longitudinal and the transverse apertures. The paper compares two approaches to select the optimal sextupole strengths. The first one is multi-objective Bayesian optimization, where the dynamic aperture volume from tracking simulations is used as an objective to be maximized. The second approach does not involve tracking and minimizes the geometric and chromatic resonance driving terms. The comparison of the two results includes their quality in terms of the size of the achievable 3D dynamic aperture and the computational effort involved.

This paper addresses the automation of particle accelerator control through reinforcement learning (RL). It highlights the potential to increase reliable performance, especially in light of new diagnostic tools and the increasingly complex variable schedules of specific accelerators. We focus on the physics simulation of the AWAKE electron line, an ideal platform for performing in-depth studies that allow a clear distinction between the problem and the performance of different algorithmic approaches for accurate analysis. The main challenges are the lack of realistic simulations and partially observable environments. We show how effective results can be achieved through meta-reinforcement learning, where an agent is trained to quickly adapt to specific real-world scenarios based on prior training in a simulated environment with variable unknowns. When suitable simulations are lacking or too costly, a model-based method using Gaussian processes is used for direct training in a few shots only. The work opens new avenues for implementing control automation in particle accelerators, significantly increasing their efficiency and adaptability.

Reinforcement Learning (RL) has demonstrated its effectiveness in solving control problems in particle accelerators. A challenging application is the control of the microbunching instability (MBI) in synchrotron light sources. Here the interaction of an electron bunch with its emitted coherent synchrotron radiation leads to complex non-linear dynamics and pronounced fluctuations. Addressing the control of intricate dynamics necessitates meeting stringent microsecond-level real-time constraints. To achieve this, RL algorithms must be deployed on a high-performance electronics platform. The KINGFISHER system, utilizing the AMD-Xilinx Versal family of heterogeneous computing devices, has been specifically designed at KIT to tackle these demanding conditions. The system implements an experience accumulator architecture to perform online learning purely through interaction with the accelerator while still satisfying strong real-time constraints. The preliminary results of this innovative control paradigm at the Karlsruhe Research Accelerator (KARA) will be presented. Notably, this represents the first experimental attempt to control the MBI with RL using online training only.

Reinforcement Learning (RL) is a unique learning paradigm that is particularly well-suited to tackle complex control tasks, can deal with delayed consequences, and can learn from experience without an explicit model of the dynamics of the problem. These properties make RL methods extremely promising for applications in particle accelerators, where the dynamically evolving conditions of both the particle beam and the accelerator systems must be constantly considered. While the time to work on RL is now particularly favorable thanks to the availability of high-level programming libraries and resources, its implementation in particle accelerators is not trivial and requires further consideration. In this context, the Reinforcement Learning for Autonomous Accelerators (RL4AA) international collaboration was established to consolidate existing knowledge, share experiences and ideas, and collaborate on accelerator-specific solutions that leverage recent advances in RL. Here we report on two collaboration workshops, RL4AA'23 and RL4AA'24, which took place in February 2023 at the Karlsruhe Institute of Technology and in February 2024 at the Paris-Lodron Universität Salzburg.

ROCK-IT aims to develop a demonstrator for automation and remote-access to beamlines of synchrotron radiation facilities. The four participating Helmholtz centers DESY, HZB, HZDR, and KIT have identified catalysis operando experiments as a pilot development. So far, no automation exists for such experiments and since the optimization of catalysts requires to evaluate a large parameter space of experimental and material conditions, it is a perfect demonstrator case for a prototype. For the research community, a suitable automation of such experiments will allow for a more effective development workflow. For KIT's catalyze beamline CAT at the Karlsruhe Research Accelerator KARA a prototype of the setup is currently in development.

Understanding the evolution of complex systems with numerous interacting particles requires advanced analytical tools capable of capturing the intricate dynamics of the phase space. This study introduces a novel approach to longitudinal phase space density tomography in an electron storage ring, leveraging constraints imposed by the Vlasov-Fokker-Planck equation. The Vlasov-Fokker-Planck equation provides a comprehensive description of the evolution of density functions in phase space, accounting for both deterministic and stochastic processes. Measurements of the turn-by-turn bunch profile offer a time-dependent projection of the phase space. Observing the bunch profile evolution of charged particles in regimes characterized by a rich phase space dynamics presents a challenging inverse problem for reconstructing the phase space densities. In this work, we present a tomographic framework for reconstructing the longitudinal phase space density of an electron bunch at the Karlsruhe Research Accelerator (KARA). This framework utilizes simulated data and applies the Vlasov-Fokker-Planck equation to drive the reconstruction process.