Hi Irek,

On 22 November 2015 at 19:11, Ireneusz Szcześniak <irek.szczesniak@gmail.com> wrote:

Hi Jeremy and others,

I implemented the edge-disjoint KSP, but without a min-cost max-flow algorithm. You can find the implementation here:

https://svn.boost.org/trac/boost/ticket/11804

That's great, it looks good. I don't have time to take a detailed look at it right now, but at first glance, I'm wondering whether you really have to use multimap, set and list? Have you benchmarked against a different implementation using a data structure with better locality of storage (e.g. vector)?

Jeremy, did you try to implement Yen? I think I'll have a stab at it.

No, not yet, I've been busy with polynomial division in Boost.Math. I probably won't really have time to look at it for a few weeks.

Cheers.

Jeremy