GeOFF is a modular Python framework harmonizing optimization and reinforcement learning approaches for accelerator tuning.
It provides standard interfaces for optimization problems, adapters to third-party solvers, and a graphical frontend (geoff-app) for operator interaction.
Developed at CERN and GSI under the EURO-LABS program.
Paper: “GeOFF: The Generic Optimization Framework & Frontend for Particle Accelerator Controls” (arXiv 2506.03796)
Repository: gitlab.cern.ch/geoff/geoff-app
GeOFF
Generic Optimization Framework & Frontend for Particle Accelerator Controls