The fluid dynamic approach to equidistribution methods for grid generation and adaptation
- Los Alamos National Laboratory
The equidistribution methods based on L{sub p} Monge-Kantorovich optimization [Finn and Delzanno, submitted to SISC, 2009] and on the deformation [Moser, 1965; Dacorogna and Moser, 1990, Liao and Anderson, 1992] method are analyzed primarily in the context of grid generation. It is shown that the first class of methods can be obtained from a fluid dynamic formulation based on time-dependent equations for the mass density and the momentum density, arising from a variational principle. In this context, deformation methods arise from a fluid formulation by making a specific assumption on the time evolution of the density (but with some degree of freedom for the momentum density). In general, deformation methods do not arise from a variational principle. However, it is possible to prescribe an optimal deformation method, related to L{sub 1} Monge-Kantorovich optimization, by making a further assumption on the momentum density. Some applications of the L{sub p} fluid dynamic formulation to imaging are also explored.
- Research Organization:
- Los Alamos National Laboratory (LANL), Los Alamos, NM (United States)
- Sponsoring Organization:
- USDOE
- DOE Contract Number:
- AC52-06NA25396
- OSTI ID:
- 962380
- Report Number(s):
- LA-UR-09-01688; LA-UR-09-1688; TRN: US200920%%136
- Journal Information:
- SIAM Journal of Scientific Computing, Journal Name: SIAM Journal of Scientific Computing
- Country of Publication:
- United States
- Language:
- English
Similar Records
Robust, multidimensional mesh-motion based on Monge-Kantorovich equidistribution
Robust, multidimensional mesh motion based on Monge-Kantorovich equidistribution