Seminars

NONLINEAR MULTIDIMENSIONAL MODELLING IN SCIENTIFIC COMPUTING

by Prof. dr. Annie Cuyt, RESEARCH DIRECTOR FWO
Mathematics & Computer Science
Universiteit Antwerpen (CMI)

In  Byron 516 at 11:15 on Friday January 28th


Nonlinear approximations have proved extremely useful in studying several industrial problems. We show how results from rational approximation theory can be applied to

(*) filter design,
(*) metamodelling in microwave theory,
(*) computer graphics,
(*) financial mathematics,
(*) shape reconstruction and
(*) biomedical signal processing.

Each problem may supply data of a slightly different nature: grid structured or scattered, accurate or subject to uncertainty, bountiful or sparse. We indicate how theoretical results on convergence behaviour and error estimation lead to useful rules of thumb in practical applications. For data suffering from uncertainty and for applications requiring error control, we introduce a new interval interpolation technique.