Re: [Help-gsl] QRNG

[Brian Gough]

Philipp Noël Baecker writes:
 > I adapted a QuantLib function known to be suitable for that purpose.
 > However, there seems to be no improvement over gsl_cdf_ugaussian_Pinv in the
 > simulation result.
 > I would be grateful for any suggestions.

The algorithm used in gsl_cdf_ugaussian_Pinv (cdf/gaussinv.c) is very
similar.  I assume that you use it in combination with a underlying
QRNG for the [0,1] sequence.

 > PS: It is inconvenient that PRNG and QRNG have different datatypes so that
 > functions have to be customized to accept one or the other. Is there a way
 > to use a "superclass" for this purpose?

It should be fairly straightforward to write a QRNG generator which is
purely random, but it is difficult to go the other way because the RNG
methods do not know anything about dimension.

-- 
Brian Gough

Network Theory Ltd
http://www.network-theory.co.uk/