[pgrouting-users] OpenGraphRouter

Stephen Woodbridge woodbri at swoodbridge.com
Mon Mar 7 00:33:20 EST 2011


Hi Jay,

Regarding OpenGraphRouter, the idea was to start a project to create of 
open source routing engine for navigation and possibly other graph 
analysis. Our goal is to build a commercial grade routing engine capable 
of find routes through 25-30 million node networks in 1 second ball park 
times.

An important part of this project long term was to be able to use the 
library in multiple projects like pgRouting, as a standalone webservice 
by writing a cgi-bin or fast-cgi handler interface for it. There is also 
interest from Ingres database to integrate such a library.

For a lot of reasons we wanted to license this similar to the 
Mapserver.org project which is MIT-X style license. The thought behind 
this is that we could potentially get funding down the road from 
companies that want features. And many of the potential companies will 
not deal with GPL licenses.

I wrote a router 10 years ago as a prototype and figured out how package 
up various tools, like TSP, drive time analysis, given a fleet of 
vehicles the closest one to an arbitrary point like a package pickup so 
you could say divert that truck to the point. I have also managed a 
commercial grade router in a production environment and have a really 
good idea on how things work and what is needed for this. But I do not 
have the time to focus on writing the tools we need.

So I mentored a student from Bangladesh for the last two years during 
Google Summer of code and his code is now in the OpenGraphRouter SVN. 
This last year, I was very busy and do not feel like I managed the 
project very well as SVN is a little chaotic because things got checked 
into the wrong places and it has not been cleaned up.

At this point if someone wants to work on developing this I would be 
happy to work with them and the pgRouting team to try and push this 
forward. This has some advantages in that building a library if we 
adhere to some back rules it is probably easier to develop and debug, 
and then we can move it into the pgrouting environment. Also being a 
library means the we can get more exposure to the code base which I 
think will help improve our ability to get talent qualified to support 
the ongoing development and hopefully build some critical mass.

Anyway that is the idea.

Best regards,
   -Steve


More information about the Pgrouting-users mailing list