Abstract:
There has been enormous progress made over the last decade in developing
fast-multipole accelerated integral equation solvers for Maxwell's equations in
both the low and high frequency regimes. In order to create a supportable
software infrastructure, however, issues of FMM acceleration, geometry
representation, quadrature and iterative solution strategy should be cleanly
separated. We will describe some of our recent efforts in this direction as well
as a new class of direct solvers which have been designed to effectively model
nearly resonant structures.
[LECTURE
SLIDES]
|