BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//talks.bham.ac.uk//v3//EN
BEGIN:VEVENT
CATEGORIES:Applied Mathematics Seminar Series
SUMMARY:A posteriori error estimation for stochastic Galer
kin approximations - Alex Bespalov (Birmingham)
DTSTART:20140310T140000Z
DTEND:20140310T150000Z
UID:TALK1354AT
URL:/talk/index/1354
DESCRIPTION:Stochastic Galerkin finite element method is an in
creasingly popular approach for the solution of el
liptic PDE problems with correlated random data. I
t combines conventional (*h-*) finite element
approximation on the spatial domain with spectral
(*p-*) approximation on a finite-dimensional
manifold in the (stochastic) parameter space. Ada
ptive techniques are now under development to capt
ure this finite-dimensional manifold and to constr
uct efficient spatial and stochastic approximation
s.\n\nIn this talk\, we outline the issues involve
d in a posteriori error analysis of computed solut
ions and present a practical a posteriori estimato
r for the approximation error. We also discuss dif
ferent strategies for enriching the discrete space
and derive computable estimates of the error redu
ction for the corresponding enhanced approximation
s. We show numerically that these estimates can be
used to choose the enrichment strategy that reduc
es the error most efficiently\, and thus they can
guide an adaptive refinement algorithm.
LOCATION:Watson LRA
CONTACT:Alexandra Tzella
END:VEVENT
END:VCALENDAR