[SCIP] SCIPchgReoptObjective() without resetting other variable coeffs

[Christopher G]

Dear Ambros,

thank you for your comprehensive answer. This was very insightful.

Best,
Christopher

------- Original Message -------
Ambros Gleixner <gleixner at zib.de> schrieb am Freitag, 14. Juli 2023 um 13:29:


> Hi Christopher,
> 
> No, but you are correct that this is a good feature request. It is a
> bit tricky to implement (but not impossible) because of the link between
> original and transformed variables.
> 
> If you are worried about overhead for memory allocation, this should be
> solved by using SCIPallocCleanBufferArray(), see
> 
> https://scipopt.org/doc/html/group__PublicMemoryMethods.php#ga11c9a6e3b9eb81c15d84e2ef636c40d0
> 
> and
> 
> https://scipopt.org/doc/html/MEMORY.php
> 
> Important: You have to reset the changed entries of the array to zero
> before freeing it.
> 
> Best,
> Ambros
> 
> 
> Am 12.07.2023 um 17:06 schrieb Christopher G:
> 
> > Good day SCIP community,
> > 
> > currently I try to use the reoptimization feature of SCIP to
> > accelerate the consecutive solving process of my pricing problems in a B&P
> > algorithm. I was wondering whether the method SCIPchgReoptObjective(..) has a
> > counterpart which can be used to adapt the objective function coefficients for a
> > set of variables while not resetting the coefficients of all other variables.
> > 
> > Since I don't have contiguous memory for all the coefficients of all my
> > variables in my pricing problem, I have to allocate a (rather big) chunk of
> > memory that contains such coefficients in each iteration of my pricing. This is
> > rather cumbersome and I'd like to avoid this.
> > 
> > Best,
> > Christopher
> > 
> > _______________________________________________
> > Scip mailing list
> > Scip at zib.de
> > https://listserv.zib.de/mailman/listinfo/scip
> 
> _______________________________________________
> Scip mailing list
> Scip at zib.de
> https://listserv.zib.de/mailman/listinfo/scip