Přejít k obsahu


Bayesian optimization methodology for design optimization of electromagnetic system

Citace:
MACH, F., KARBAN, P., DOLEŽEL, I., ŠROUBOVÁ, L., ŠMÍDL, V. Bayesian optimization methodology for design optimization of electromagnetic system. In ESCO 2016 : 5th European Seminar on Computing : Book of Abstracts. Pilsen: University of West Bohemia in Pilsen, 2016. s. 162., ISSN: neuvedeno
Druh: STAŤ VE SBORNÍKU
Jazyk publikace: eng
Anglický název: Bayesian optimization methodology for design optimization of electromagnetic system
Rok vydání: 2016
Místo konání: Pilsen
Název zdroje: University of West Bohemia in Pilsen
Autoři: Ing. František Mach Ph.D. , Doc. Ing. Pavel Karban Ph.D. , Prof. Ing. Ivo Doležel CSc. , Ing. Lenka Šroubová Ph.D. , Doc. Ing. Václav Šmídl Ph.D.
Abstrakt EN: Mathematical optimization is pervasive in the most of the engineering problems. From practical point of view, real problems are computationally very expensive and enormous effort has to be exerted to find the best available values of the set of objective functions. Apart from the complexity of the optimization problem, huge attention has to be paid to address uncertainty in the optimization process. Nowadays, one of the promising optimization techniques is the Bayesian optimization, which seem to be the cutting edge in the current state-ofthe-art approaches. The fundamentals of this methodology is presented and discussed from general viewpoint. All theoretical assumptions are verified by numerical experiments with the artificial test functions and also tested on practical engineering optimization problems.
Klíčová slova

Zpět

Patička