[Cbc] Fw: cbc MILP 'warm start'

[Gabrielle A. Grün]

----- Original Message ----- 
From: "Gabrielle A. Grün" <grun at cs.sfu.ca>
To: "Matthew Galati" <magh at lehigh.edu>
Sent: Tuesday, June 03, 2008 2:17 PM
Subject: Re: [Cbc] cbc MILP 'warm start'


>
>
> Hi Matt,
>
> For 2, look at setting the cutoff in CbcModel; watch for the offset 
> bestObjective_  is key too.  Thanks.
>
>
>
>
> Gabrielle A. Grün, Ph.D. Student
> School of Computing Science
> Simon Fraser University
> 8888 University Drive
> Burnaby, BC
> V5A 1S6
> <http://www.cs.sfu.ca/~grun>
> ----- Original Message ----- 
> From: "Matthew Galati" <magh at lehigh.edu>
> To: <cbc at list.coin-or.org>; <osi at list.coin-or.org>
> Sent: Tuesday, June 03, 2008 1:17 PM
> Subject: [Cbc] cbc MILP 'warm start'
>
>
>>A few questions:
>> (1) Can CBC be 'warm-started' with a full and/or partial MILP solution?
>> For things like red-cost fixing, fathoming, etc. If so, how?
>> (2) Can CBC be given a starting upper bound (for a minimization)? i.e.,
>> (1) w/o the actual solution, just the bound.
>> (3) Can (1) and (2) be done through the OsiCbc / any Osi interface?
>> (4) I am pretty sure Cplex can do (1) and (2), how would one do this
>> through the OsiCpx interface?
>> (5) Where can I find a list of Cbc (library) parameters and what they
>> do? I suspect, if (2) is possible, it is just a parameter - but I am not
>> sure where I find that - I checked CbcParam.hpp but there are no
>> descriptions.
>>
>> Thanks,
>> Matt
>>
>> _______________________________________________
>> Cbc mailing list
>> Cbc at list.coin-or.org
>> http://list.coin-or.org/mailman/listinfo/cbc
>>
>