It contains universal (also called automatic or black-box) algorithms that can generate random numbers from large classes of continuous or discrete distributions, and also from practically all standard distributions.

to generate random numbers the user must supply some information about the desired distribution, especially a c-function that computes the density and - depending on the chosen methods - some additional information (like the borders of the domain, the mode, the derivative of the density ...). after a user has given this information an init-program computes all tables and constants necessary for the random variate generation. the sample program can then generate variates from the desired distribution.

this package contains the development files

root web-site: http://root.cern.ch unuran web-site: http://statistik.wu-wien.ac.at/unuran/