[Scip] Implementing new branching rule

Timo Berthold berthold at zib.de
Tue Dec 18 22:53:10 MET 2012


Well, for each variable and constraint you can get its transformed:
SCIP{var,cons}GetTransformed()

You can easily look in the LP/var/cons section of scip.h and into
pub_var.h, pub_cons.h, pub_lp.h to see whether the interface functions
that you need is available there.

Timo

> Actually thinking a bit more, what I actually need is the transformed
> problem constraints and the variables in each constraint. Is there a way
> to
> access those?
>
> Thanks
> --Sid
>
>
> On Tue, Dec 18, 2012 at 3:42 PM, Siddhartha Jain <sj1 at cs.cmu.edu> wrote:
>
>> Hi Timo,
>>
>> Thanks for the reply. It's a research idea that I'm implementing.
>>
>> If the variables returned are those of the transformed problem, is there
>> a
>> way for me to get the column index to which the variables correspond to?
>>
>> Thanks
>> --Sid
>>
>>
>> On Tue, Dec 18, 2012 at 3:16 PM, Timo Berthold <berthold at zib.de> wrote:
>>
>>> Hi Sid,
>>>
>>> this function gives you an array of SCIP_VAR pointers. These variables
>>> are
>>> variables in the transformed problem which have a fractional solution
>>> in
>>> the LP solution of the current node relaxation.
>>>
>>> If you want to get the variable's problem index, you can call the
>>> function
>>> SCIPvarGetProbindex() with these variable pointers as input.
>>>
>>> I am curious: what variant of strong branching do you plan to
>>> implement?
>>>
>>> I recommend you to look through the full strong branching and maybe the
>>> most fractional branching rule in SCIP to get an idea how they are set
>>> up.
>>>
>>> If not yet happened, I further recommend you to have a look at the
>>> following points in the SCIP documentation:
>>> http://scip.zib.de/doc/html/BRANCH.shtml
>>>
>>> http://scip.zib.de/doc/html/scip_8h.shtml#a04924d1578c8b893f8420001445210fd
>>>
>>> http://scip.zib.de/doc/html/pub__var_8h.shtml#a661c5d1f7f4e09471f8d2898b5bda91c
>>>
>>> Also check out our FAQ.
>>>
>>> Best regards,
>>> Timo
>>> > Hi,
>>> >
>>> > I'm implementing a variant of strong branching within SCIP. I was
>>> > wondering
>>> > - what do the lpcands obtained via the function SCIPgetLPBranchCands
>>> > correspond to?
>>> >
>>> > Do they correspond to the original problem variables or the
>>> constraint
>>> > matrix variables after presolving, etc.
>>> >
>>> > How can I obtain the problem variable index/column index resp. the
>>> > variable
>>> > is referring too?
>>> >
>>> > Thanks
>>> > --Sid
>>> > _______________________________________________
>>> > Scip mailing list
>>> > Scip at zib.de
>>> > http://listserv.zib.de/mailman/listinfo/scip
>>> >
>>>
>>>
>>
>



More information about the Scip mailing list