[boost] Re: Hypergraphs and the BGL

From: Jeremy Siek (jsiek_at_[hidden])
Date: 2002-04-26 08:33:23

Hi Chris,

It would not be easy to represent a hypergraph with existing BGL classes
or interfaces. If you search in the boost mailing list archives you will
find some earlier discussions about hypergraphs. There was someone 
interested
in adding hypergraph support to the BGL, but that work was not finished.
The interface I suggested for hypergraphs would be that a hypergraph would
not use the source() and target() functions of the current BGL interface,
and instead would have an incident_vertices() function that returns an
iterator range.

Cheers,
Jeremy

--On Friday, April 26, 2002 7:08 AM -0500 boost-request_at_[hidden] 
wrote:
> Can a hypergraph be represented atomically by the BGL? Or does it make
> more sense to think of a generic hypergraph container, and generic
> hypergraph algorithms implemented on top of the BGL? I'm trying to bust a
> problem that lends itself to hypergraph analysis and finding that it
> won't yield. Contemplating bringing the mountain to it. Any thoughts on
> this matter would be appreciated.
>
> - Regards
> Chris Russell