Title :
Gradient-Based Worst Case Search Algorithm for Robust Optimization
Author :
Chiariello, Andrea G. ; Formisano, Alessandro ; Martone, Raffaele ; Pizzo, Francesco
Author_Institution :
Dept. of Ind. & Inf. Eng., Seconda Univ. di Napoli, Caserta, Italy
Abstract :
The worst case search plays a fundamental role in the design of electromagnetic devices when robustness versus tolerances has to be pursued. Here, an approach is proposed to evaluate in effective and fast way the worst solution, when both the performance function and tolerance constraint function gradients are available. The algorithm is then generalized for the cases when such information is not explicitly given.
Keywords :
design engineering; electromagnetic devices; gradient methods; optimisation; search problems; electromagnetic device design; gradient-based worst case search algorithm; performance function gradient; robust optimization; tolerance constraint function gradient; Approximation methods; Estimation; Linear programming; Magnetic resonance imaging; Optimization; Robustness; Superconducting magnets; Robust design; tolerance analysis; worst case;
Journal_Title :
Magnetics, IEEE Transactions on
DOI :
10.1109/TMAG.2014.2356798