Reduced Order Models (POD) for Calibration Problems in Finance

The text is written in December 2008 for the http://groups.google.com/group/mor4ansys group.

E. W. Sachs and M. Schu
Reduced Order Models (POD) for Calibration Problems in Finance
Numerical Mathematics and Advanced Applications, 2008, Part 2, p. 735-742
http://dx.doi.org/10.1007/978-3-540-69777-0_88

It is exiting to learn that model reduction could help to save us from the financial crisis.

In the paper they apply POD to some nonlinear partial integro-differential equation. I should confess that I have not understood the equation by itself, it is quite a different topic from PDEs that I was working with.

What is interesting that they have been ably to use a reduced model for an optimization, at least partly. If I have understood correctly, they have computed a POD basis for some chosen parameters but then while projecting they have preserved some parameters. It is also working for normal model reduction provided the range of parameters is small.


Comments are closed.