Evolution and Genetic Optimization of Higher-Order Shim Coils for NMR

0205338 - UPT-D 20000117 RIV CZ eng C - Konferenční příspěvek (zahraniční konf.)
Chládek, Jan - Konzbul, Pavel - Ošmera, P.
Evolution and Genetic Optimization of Higher-Order Shim Coils for NMR.
Mendel 2000 - 6th International Conference on Soft Computing. Brno: Brno University of Technology, Faculty of Mechanical Engineering, 2000, s. 39-44. ISBN 80-214-1609-2.
[Mendel 2000 /6./ - Soft Computing. Brno (CZ), 07.06.2000-09.06.2000]
Grant CEP: GA MŠMT ME 181; GA ČR GA102/95/0467
Výzkumný záměr: CEZ:AV0Z2065902
Kód oboru RIV: BM - Fyzika pevných látek a magnetismus

Evolutionary and genetic scholastic optimizations were explored for designing NMR gradient and shim coils. Some hierarchical self-adaptive algorithms, based on a concept of meta-optimization, were tested. Time-demands and quality of the solutions were compared for various optimisers. A broad range of contrasting features was found for different settings of governing parameters and for different constraint handling's. Therefore a fuzzy logic in meta-decision level and an adaptative barrier function were used for optimization of multi-modal objective function. The using of meta-level optimization of different parameters was identified as the major challenge for future investigations and improvements of the optimizers.
Trvalý link: http://hdl.handle.net/11104/0100952