I've just downloaded quantlib and started playing around with it, and it looks like it's designed primarily to use Euler discretizations for everything -- so far as I can tell, there's not even a method provided to exactly simulate geometric Brownian motion. Am I missing an obvious feature? If not, is there a fundamental reason why you wouldn't (ever) want to use exact numerics instead of approximate ones?
$\begingroup$
$\endgroup$
Add a comment
|
1 Answer
$\begingroup$
$\endgroup$
Well, actually it's designed to use whatever discretization you throw at it---but for the time being only Euler discretization is implemented, mostly for lack of time or interest on the part of contributors. If you want to use exact numerics with a process, you can just code the corresponding discretization class (you'll have to inherit from StochasticProcess::discretization) and pass an instance of your class to the process upon instantiation. The whole design is described in more detail in chapter 6 at http://implementingquantlib.blogspot.com/p/the-book.html, or rather, the half of chapter 6 I've written so far.
Needless to say, if you do write exact discretization, send a patch our way and we'll include it in the library.
