The Joint Accelerator Conferences Website (JACoW) is an international collaboration that publishes the proceedings of accelerator conferences held around the world.
TY - CONF AU - Appel, S. AU - Reimann, S. ED - Boland, Mark ED - Tanaka, Hitoshi ED - Button, David ED - Dowd, Rohan ED - Schaa, Volker RW ED - Tan, Eugene TI - Beam Line Optimization Using Derivative-Free Algorithms J2 - Proc. of IPAC2019, Melbourne, Australia, 19-24 May 2019 CY - Melbourne, Australia T2 - International Particle Accelerator Conference T3 - 10 LA - english AB - The present study focuses on the beam line optimization from the heavy-ion synchrotron SIS18 to the HADES experiment. BOBYQA (Bound Optimization BY Quadratic Approximation) solves bound constrained optimization problems without using derivatives of the objective function. The Bayesian optimization is an other strategy for global optimization of costly, noisy functions without using derivatives. A python programming interface to MADX allow the use of the python implementation of BOBYQA and Bayesian method. This gave the possibility to use tracking simulation with MADX to determine the loss budget for each lattice setting during the optimization and compare both optimization methods. PB - JACoW Publishing CP - Geneva, Switzerland SP - 2307 EP - 2310 KW - experiment KW - target KW - heavy-ion KW - site KW - interface DA - 2019/06 PY - 2019 SN - 978-3-95450-208-0 DO - DOI: 10.18429/JACoW-IPAC2019-WEPMP005 UR - http://jacow.org/ipac2019/papers/wepmp005.pdf ER -