Dear Colleagues,
first of all let me apologize with Sandor, as I reported his
private message on the list, instead of cutting it. I thought
his questions could be of general interest.
On Tue, 20 Jun 2000 [log in to unmask] wrote:
> Dear Colleagues
[...]
>
> There are a few more rare FX option types
> which do require simulation or more complex techniques, but from what
> I can make out from the two posts, it seems they are inventing a new
> method to do something which was not difficult in the first place! It
> only requires a calculator to price most FX options. Of course, if
> the circuit generates interesting self-similar distributions, that
> could well be of use in FX modelling techniques.
Unfortunately, I was not at the APS meeting and I did not meet Hideki
Takayasu recently. Therefore, I do not know what they exactly did.
Perhaps, they may wish to build a calculator able to cover exchange risk
when the underlying rate follows a non-Gaussian leptokurtic distribution.
By the way, using dedicated computers for Monte Carlo simulations
indeed helped in saving money and improving computational speed. Usually,
the flexibility of the general-purpose computers was lost. I guess that
special-purpose computers were more fashionable in the '80s, when
computing resources where scarcer than today. Here are some
references on dedicated computers:
http://xxx.lanl.gov/abs/hep-lat/9611014
http://www.winlab.rutgers.edu/~ato/projects.html#past
In Italy, a research project on parallel computers for computational
physics was still running until recently (I have no recent news on the
state of the project):
http://chimera.roma1.infn.it/ape.html
> >
> > Jessica
> >
> > ---
[...]
Regards
Enrico
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|