Speaker
Description
We present the next-generation version of GIOTTO, a genetic-algorithm (GA) based optimization software designed for complex beam-physics and RF-engineering problems. The upgraded framework, recently funded by INFN as the GIOTTO Project, couples GA optimization with a flexible software-orchestration layer capable of driving multi-physics tools such as ANSYS, BuildCavity (Superfish), Geant4, and arbitrary Python-controlled simulation tools. This enables multi-objective cross-disciplinary optimization of beam dynamics, RF cavity geometry, and engineering constraints within a unified workflow.
A major ongoing development is the integration of GIOTTO into EPICS-based control systems, allowing seamless switching between simulation “digital-twin” models and online accelerator optimization.
To improve convergence, GIOTTO will incorporate GPU-accelerated surrogate modeling and predictive sampling using an NVIDIA RTX 6000 Blackwell workstation GPU. These machine learning (ML) assisted strategies are expected to reduce optimization time by up to one order of magnitude while improving exploration of high-dimensional parameter spaces.
This work demonstrates a general, extensible framework to bring advanced optimization, ML, and real-time control to modern accelerator facilities.
| In which format do you inted to submit your paper? | LaTeX |
|---|