Error analysis of meshfree approximation methods based on optimization

Start: 11/18/2010 - 4:15pm
End  : 11/18/2010 - 5:15pm

Statistics/OR/Math Finance Seminar

Agustin Bompadre, CalTech


Approximation methods are used to solve numerically PDEs in Material Science. Finite Element Methods are the most usual technique used in this context. In this talk, we discuss meshfree approximation methods, which are approximation methods that do not need a meshing of the domain. We give a set of sufficient conditions to guarantee the convergence of approximates based on meshfree methods. We review the Local Maximum-Entropy approximation scheme, which is a meshfree approximation method based on an optimization principle, and we show that the general error bounds apply to this method.

Harvey Mudd College 3rd floor Sprague. Refreshments at 4pm.