While trying to implement the algorithm for gamma distributed random numbers (pp. 292-293) I was missing a second parameter needed to describe the gamma distribution:

f(x)= (b^p)/G(p) * x^(p-1)exp(-bx)

(p = "alpha" or "a" and b = 1/"beta" in angloamerican literature)

[http://de.wikipedia.org/wiki/Gammaverteilung]
[http://en.wikipedia.org/wiki/Gamma_distribution]

In the algorithm the parameter b seems to be set to 1.

Is there any posibility to change the algorithm to handle this second parameter?

Thanks in advance for your help!


/Elandril