quasi random no's.

[Info-1446]

"I've a question re: the random no. generators used in
quant lib.. for the monte carlo pricer don't you need
to use some low discrepancy (sobol) sequences,
otherwise anything with dimensions >1 is likely to be
inaccurately priced ?? numerical recipies has an
algorithm, but its limited to 6 dimensions. Anyone
know a good algorithm for going beyond this?"

Yes, there is such algorithm. Please, check http://www.broda.co.uk/software.htm

Regards
Sergei

[Michael Lewis-5]

Thanks for the information.

Rgds

Mike

--- Info <[hidden email]> wrote:

> "I've a question re: the random no. generators used
> in
> quant lib.. for the monte carlo pricer don't you
> need
> to use some low discrepancy (sobol) sequences,
> otherwise anything with dimensions >1 is likely to
> be
> inaccurately priced ?? numerical recipies has an
> algorithm, but its limited to 6 dimensions. Anyone
> know a good algorithm for going beyond this?"
>
> Yes, there is such algorithm. Please, check
> http://www.broda.co.uk/software.htm
>
> Regards
> Sergei
>
>
> _______________________________________________
> Quantlib-users mailing list
> [hidden email]
>

https://lists.sourceforge.net/lists/listinfo/quantlib-users


__________________________________________________
Do You Yahoo!?
Yahoo! Sports - live college hoops coverage
http://sports.yahoo.com/

[Peter Schmitteckert-2]

Salut,

On Friday 15 March 2002 18:23, Info wrote:

> "I've a question re: the random no. generators used in
> quant lib.. for the monte carlo pricer don't you need
> to use some low discrepancy (sobol) sequences,
> otherwise anything with dimensions >1 is likely to be
> inaccurately priced ?? numerical recipies has an
> algorithm, but its limited to 6 dimensions. Anyone
> know a good algorithm for going beyond this?"
>
> Yes, there is such algorithm. Please, check
> http://www.broda.co.uk/software.htm
>

in case you're satisfied with 40 dimensions, take a look as my
homepage:
http://www.schmitteckert.com/LinuxMagazin/Sobol/

there is a Sobol.tar.gz wich include sources for up to 40 dimensional Sobol
sequences. For those who understand german, there is a small article
in a last year "Linux Magazin" by myself.

If you want to include this to Quantlib I would change the license to
whatever fits to QuantLib.

Best wishes,
Peter

P.S.
I had problems posting to Quantlib-users complainig that
schmitteckert.org has no postmaster, but I was posting from schmitteckert.com
which indeed has a postmaster account.

[Ferdinando M. Ametrano-2]

>in case you're satisfied with 40 dimensions, take a look as my
>homepage:
>http://www.schmitteckert.com/LinuxMagazin/Sobol/
>
>there is a Sobol.tar.gz wich include sources for up to 40 dimensional Sobol
>sequences. For those who understand german, there is a small article
>in a last year "Linux Magazin" by myself.
>
>If you want to include this to Quantlib I would change the license to
>whatever fits to QuantLib.

Thank you Peter. I would like to include it in QuantLib: the QuantLib
license is a BSD modified license, please see http://quantlib.org/license.html
If the license it's OK for you tere are no problems. BTW you will maintain
the copyright and  so will be listed along with the others QuantLib
copyright holders.
Would you mind proving english comments and maybe also an english summary
of the Linux Magazine article that I can include in the QuantLib documentation?

[Peter Schmitteckert-2]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Dear Ferdinando,

On Tuesday 19 March 2002 16:08, Ferdinando Ametrano wrote:

> >in case you're satisfied with 40 dimensions, take a look as my
> >homepage:
> >http://www.schmitteckert.com/LinuxMagazin/Sobol/
> >
> >there is a Sobol.tar.gz wich include sources for up to 40 dimensional
> > Sobol sequences. For those who understand german, there is a small
> > article in a last year "Linux Magazin" by myself.
> >
> >If you want to include this to Quantlib I would change the license to
> >whatever fits to QuantLib.
>
> Thank you Peter. I would like to include it in QuantLib: the QuantLib
> license is a BSD modified license, please see
> http://quantlib.org/license.html If the license it's OK for you tere are no
> problems. BTW you will maintain the copyright and  so will be listed along
> with the others QuantLib copyright holders.

Herewith I grant you to use my sobol sources under the QuanlibLicense
specified in http://quantlib.org/license.html. Please feel free to adjust the
copyright notice in the files.

> Would you mind proving english comments and maybe also an english summary
> of the Linux Magazine article that I can include in the QuantLib
> documentation?

In case you're satisfied with some short abstract I could write something
after easter.

Best wishes,
Peter

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.4 (GNU/Linux)
Comment: For info see http://www.gnupg.org

iEYEARECAAYFAjyXY6AACgkQHn79UhrElrFRPQCgu0O/8ULMnj0eQ9pZnYSv5cWs
6rkAoLPQunnbqBRn+eHBGcVSBiCD9CJ2
=Tvc8
-----END PGP SIGNATURE-----