Thursday, October 5, 9:30 am in room MTH 3206, University of Maryland,
College Park
Specialized numerical methods for time dependent
partial differential equations
Prof. Donald French
Department of Mathematics,
University of Cinncinatti and University of Maryland Baltimore County
french@ucunix.san.uc.EDU
Accurate representation of the qualitative aspects of partial
differential equation (pde) problems is useful and often crucial in
the calculation of a sensible numerical approximation. We look at
several physical models where this approach has been successfully
applied, highlighting important properties of the pde and the way in
which the numerical method retains them. We will also discuss a more
general approach called the continuous Galerkin (cG) method which
is a technique to develop time discretizations using finite elements
and is similar to the discontinuous Galerkin (dG) method. The cG
method retains energy properties for pde models and, like the dG
method, provides a framework for the derivation of a posteriori
estimates (useful for the creation of adaptive schemes) and analysis
of the accuracy of the approximation.
|