[graph] Are you interested in the Route Inspection Problems and Vehicle Routing Problems?
Are you interested in implementations for the following problems: - Route Inspection Problems: http://en.wikipedia.org/wiki/Route_inspection_problem - classic: find shortest path for undirected graph possibly repeating edges. O(n^3). If Eulerian path exists, it is the min. - mixed graphs. NP-complete. AKA Chinese postman problem. - k-Chinese postman problem. N postmen. - Vehicle routing problem: http://en.wikipedia.org/wiki/Vehicle_routing_problem Generalization of Route Inspection. NP-complete. Many variants: edges have one ore more weights, and multiple constraints are possible: - carrying capacity of each truck - time to travel an edge and delivery windows for each edge All of those problems have extensive literature, and obvious practical applications. Why don't you use GitHub's Issue tracker instead of the mailing list? It integrates better with the code tracking and merge requests, and is much easier to search. That would greatly reduce contribution barriers to new devs.
From: Ciro Santilli
To: boost@lists.boost.org Cc: Sent: Sunday, April 13, 2014 5:23 AM Subject: [boost] [graph] Are you interested in the Route Inspection Problems and Vehicle Routing Problems?
Are you interested in implementations for the following problems:
- Route Inspection Problems: http://en.wikipedia.org/wiki/Route_inspection_problem
- classic: find shortest path for undirected graph possibly repeating edges. O(n^3). If Eulerian path exists, it is the min. - mixed graphs. NP-complete. AKA Chinese postman problem. - k-Chinese postman problem. N postmen.
- Vehicle routing problem: http://en.wikipedia.org/wiki/Vehicle_routing_problem
Generalization of Route Inspection. NP-complete. Many variants: edges have one ore more weights, and multiple constraints are possible:
- carrying capacity of each truck - time to travel an edge and delivery windows for each edge
All of those problems have extensive literature, and obvious practical applications.
Yes, definitely.
Why don't you use GitHub's Issue tracker instead of the mailing list? It integrates better with the code tracking and merge requests, and is much easier to search. That would greatly reduce contribution barriers to new devs.
I do not know -- that is a Boost-wide decision. -- Jeremiah Willcock
On Mon, Apr 14, 2014 at 4:16 PM, Jeremiah Willcock
I do not know -- that is a Boost-wide decision.
My understanding is that there is already a tracker (TRAC: https://svn.boost.org ) so to avoid having issues spread in different places you should use it instead. Also, for code reviews and other comments, the mailing list is actually a safer archive than github and allows people to get notified using custom mail filter rules, which is more flexible than any centralized solution. I believe there have been a recent discussion about this (in the last several months).
participants (3)
-
Ciro Santilli -
Jeremiah Willcock -
Klaim - Joël Lamotte