[Coin-discuss] CBC feasibility pump heuristic

[Sebastian Nowozin]

Hello everybody,

I use CBC with some knapsack models that have additional constraints.
In the cbc source I have seen that the "feasibility pump" heuristic
for constructing feasible solutions is implemented.

Is this the original feasibilty pump by Fischetti et al.
(http://miplib.zib.de/paper/fischettigloverlodi2004.pdf), or the
improved one by Achterberg and Berthold
(http://www.zib.de/Publications/Reports/ZR-05-42.pdf) ?

For some test runs with my models I get a feasible solution very early
but it has a very high objective value.  I wonder if the improved
feasibility pump could improve this.

Thanks,
Sebastian