Tuning particle accelerators with safety constraints using Bayesian optimization

Tuning machine parameters of particle accelerators is a repetitive and time-consuming task that challenging to automate. While many off-the-shelf optimization algorithms are available, in practice their use limited because most methods do not account for safety-critical constraints each iteration, such as loss signals or step-size limitations. One notable exception safe Bayesian optimization, which data-driven tuning approach global with noisy feedback. We propose evaluate variant on two research facilities the PSI: (a) SwissFEL (b) HIPA. report promising experimental results both machines, up 16 subject 224 constraints.