reinforcement learning for magnet tuning automation
In collaboration with FRIB lab @ MSU
My research applied deep RL to streamline the tuning of particle accelerator beams at the Facility for Rare Isotope Beams at Michigan State University. By focusing on a 2-D model of the pre-separator magnet system and using simulated data, I developed an RL model that closely approximates ideal magnet settings. My work demonstrated the versatility of RL in optimizing complex systems.