Hi Martin, hi Ralf,

> Nicolas, how do you produce the elements of unlabelled combinatorial
> classes?  But somehow I guess that for Plus and Times (I.e., Union
> and Cross) it should be possible / easy.  Unfortunately, I don't
> have mupad currently...

Plus and Times are indeed done straightforwardly as in the labelled
case (it's even a little bit easier, as you don't have to manipulate
the labels). On the other hand, as soon as there is a non trivial
group G acting, things start to be more difficult. We don't do the
general case: as you mentioned, it includes for example the generation
of unlabelled graphs, which is *hard*.

For the Sets, Multisets, Cycles, Lyndon constructors, there are fast
(Constant Amortized Time) special-purpose generation algorithms. For
example the generation of Cycles and Lyndon use Joe Sawada's
algorithms which dates back from just 2003. There are also algorithms
for ranking/unranking Sets and Multisets. But just to had a little bit
of salt in the discussion: it is currently *not known* if it is
possible to efficiently unrank Cycles.

Cf. X. Molinero's thesis for an overview. 

Cheers,
                                        Nicolas

PS: Although I am quite silent, I am following what you both are
doing. This sounds really promising!

-- 
Nicolas M. ThiƩry "Isil" <[EMAIL PROTECTED]>
http://Nicolas.Thiery.name/

-------------------------------------------------------------------------
Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job easier
Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642
_______________________________________________
Aldor-combinat-devel mailing list
Aldor-combinat-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/aldor-combinat-devel

Reply via email to