July 2009


Options: Use Monospaced Font
Show Text Part by Default
Show All Mail Headers

Message: [<< First] [< Prev] [Next >] [Last >>]
Topic: [<< First] [< Prev] [Next >] [Last >>]
Author: [<< First] [< Prev] [Next >] [Last >>]

Print Reply
Sean Luke <[log in to unmask]>
Reply To:
Sat, 4 Jul 2009 16:07:37 +0200
text/plain (32 lines)
t.n.a. wrote:

> I'm not sure how I would go about transforming uniform distribution
> random numbers to a specific distribution...I could do it for simple,
> streight-lined artificial distributions, but for distributions like
> gamma, it'd turn into a project of it's own...

Actually, it's not nearly as hard as it sounds -- it depends on the 
distribution though.  There are a lot of standard algorithms (Knuth 
might be good to check out).

CERN's numerics library, which I think Repast is using, has got some 
transformations in it.  We could start there.

> Thinking about it, it seems to me that it would be good enough to
> track down one or two good libraries supporting a wide variety of
> probability distributions and do the following:
> - package them together with mason if the licences are compatible and/or
> - update the mason documentation so that it links heavily to the
> selected libraries and explains how to use them

The primary issue license wise is: we're hesitant to include LGPL 
because it has some serious compatibility problems with Java and 
BSD-style licenses (like our own).  GPL is out entirely.  So if we can 
find a BSD-style it'd be best.  Else we could just roll our own.  For 
example, I might just implement some algorithms directly in 
MersenneTwisterFast, which would be the most convenient scenario probably.

Let me know shat you find.