[Lemon-user] Lemon Cycle canceling

Jean-Bertrand Gauthier jean-bertrand.gauthier at hec.ca
Tue Aug 7 21:44:38 CEST 2012


Hello,

I'm using your library to implement a new cycle canceling algorithm.
I'm trying to figure out why the number of nodes in the residual network
is ( n + 1 ).

Also, are you aware that the solution to the circulation algorithm does
not provide an admissible primal form solution.  meaning that for
instance in your min_cost_flow_test instance, the solution given
initially contains a cycle  2-> 10 -> 4 -> 0  and 2 -> 0.  All of these
arcs have  > 0 and < u flow.  At least one of them should however be at
0 or U by modifying the cycle flow by some delta unit.




More information about the Lemon-user mailing list