RE: Masters Dissertations

Posted by Ferdinando Ametrano-3 on Apr 13, 2004; 10:42am
URL: http://quantlib.414.s1.nabble.com/Masters-Dissertations-tp2875p2878.html

Hi Daniel

>I am currently looking for dissertation topics in Mathematical
>Finance ... I would like to develop accompanying code as part of
>quantlib for whatever work I do.

great! I love your attitude: you're welcome. Niels Sønderby and Sad Rejeb
did the same for their dissertations. You've got Niels' feedback, Sad might
consider to un-lurk and provide you with his own feedback :-)

> > maybe there are areas on the Quantlib todo list I could make a
> > dissertation out of?
>
> >>I doubt it. The todo list is more concerned with implementing existing
> >>literature than with breaking new ground. Unless you can focus your
> >>thesis on implementation, but I don't think this is the case, is it?
>
>Right we do need to break some new ground - if at least
>tweaking or experimenting with some model in a new way.

the main difference is between a) using the current QuantLib code base plus
few extensions to crunch out the numbers of your thesis or b) designing
part of the QuantLib framework.


> > The areas I am particularly interested in are numerical methods in FX
> > or Credit Derivatives markets.
>
> >>Which are two areas in which QuantLib is utterly lacking. See how
> >>everything fit together neatly? :)
>
>Nice, the other model I am intersted in looking at is HJM/BGM - I think
>Would also be useful for QuantLib.

to build an HJM/BGM framework in the current QuantLib code base is a large
task and you should not embark on it unless you have plenty of time and
confidence. You might end up spending 95% of your time building the
framework and only 5% "tweaking or experimenting with it in a new way".
If this is OK with you, that's perfect for QuantLib :-)

I look forward to your contribution: larger or smaller it is welcome.

ciao -- Nando