Technical Report for Bayesian Optimization and Reinforcement Learning for Beam Polarization Increase in the BNL Hadron Injectors
Wang, Yinan · 2026
Read Technical Report for Bayesian Optimization and Reinforcement Learning for Beam Polarization Increase in the BNL Hadron Injectors by Wang, Yinan (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/doe-osti-3014455-technical-report-for-bayesian-optimizati.txt.
Description
Yinan Wang of Rensselaer Polytechnic Institute wrote this technical report for award DE-SC0024617. It covers September 2023 to August 2025. The topic is raising beam polarization in Brookhaven's hadron injectors. The tools are Bayesian optimization and reinforcement learning. Polarized beams feed RHIC and the future Electron-Ion Collider. Polarization enters the physics figure of merit to high order. Small losses thus cost real science.
The team worked with Brookhaven, Cornell, SLAC, and Jefferson Lab. They built Bayesian calibration and discrepancy models for accelerator digital twins. The models account for magnet misalignments and noisy diagnostics. Pure physics models miss those effects. A model-independent optimizer tuned the Linac-to-Booster transfer line as a black box. The conclusions judge Bayesian optimization best for low-dimensional tuning. Reinforcement learning looks promising for sequential control tasks.
Catalog record
| CREATOR | Wang, Yinan |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 48 KB |
| LICENSE | Public Domain |
| RIGHTS | PUBLIC DOMAIN |
| SOURCE | DOE OSTI |