JACoW is a publisher in Geneva, Switzerland that publishes the proceedings of accelerator conferences held around the world by an international collaboration of editors.
TY - UNPB AU - Mitsuka, G. AU - Iida, N. AU - Kato, S.K. AU - Natsui, T. AU - Satoh, M. ED - Schaa, Volker R.W. ED - Batten, Tonia ED - Bree, Michael ED - Petit-Jean-Genaz, Christine ED - Hunter, Darren TI - Machine Learning-Assisted Beam Operation at SuperKEKB and Linac at KEK J2 - Proc. of IBIC2023, Saskatoon, Canada, 10-14 September 2023 CY - Saskatoon, Canada T2 - International Beam Instrumentation Conference T3 - 12 LA - english AB - Hundreds to thousands of tuning parameters must be optimized for each operating condition to obtain the best performance from an accelerator. In the past, experts made decisions based on their experience on which tuning parameters contributed the most to the performance and adjusted them sequentially. On the other hand, accelerator tuning approaches based on machine learning, which has become much easier to handle, have been studied intensively in recent years. We have been developing a beam-tuning tool based on Bayesian optimization for boosting the SuperKEKB accelerator. In this presentation, we will report on the latest status of the beam test of the positron-beam-yield maximization and dispersion tuning at the KEK injector as the first development step. PB - JACoW Publishing CP - Geneva, Switzerland ER -