In general, this is the only way available. If you have a probability
distribution for a variable x, (i.e. p(X)dX is the probably that x
lies in [X, X + dX]), then the probability distribution for a function
of that variable, y(x), is given by p(Y) = p(X) / dy/dx. If you want
p(Y) = const, and you know that p(X) = const, then you must have dy/dx
= const => y = c + d x for some c and d. That is, only linear
functions take a uniform distribution to a uniform distribution. So,
the only way to generate a uniform distribution on [a,b] given a
uniform distribution on [0,1] is the way you outlined.
Will
On May 6, 2008, at 8:00 AM, Awhan Patnaik wrote:
GSL allows random number generation between the unit interval [0,1].
However
if one has to generate numbers between ANY two real numbers a and b
then one
way is the following:
num = a + rand*(b-a) where rand is a random number in the unit
interval. Is
this the only way available ? I mean are the statistical properties
preserved in such cases.
_______________________________________________
Help-gsl mailing list
[email protected]
http://lists.gnu.org/mailman/listinfo/help-gsl
_______________________________________________
Help-gsl mailing list
[email protected]
http://lists.gnu.org/mailman/listinfo/help-gsl
