[SCIP] Looking for example how to express symmetry groups

[Marc Pfetsch]

Dear Johann-Tobias,

the main workhorse for symmetry handling is prop_symmetry. This 
propagator stores the symmetry information that you are probably 
interested in. It stores the generators of the permutation group in 
dense format. Note that other software (like GAP or Sage) use a cycle 
notation for the permutations.

The symmetries can be output in the command line interface by "display 
symmetries.

An explanation of the symmetry handling capabilities is available in the 
SCIP Release Reports (see the SCIP webpage).

Note that SCIP can also handle symmetries on non-binary variables.

I hope that these pointers give you enough information. Otherwise, we 
would need more specific questions.

Best

Marc


On 12/06/2025 11:58, Johann-Tobias Schäg wrote:
> I am aware that there are mutliple ways to handle symmetries in SCIP
> https://www.scipopt.org/doc/html/cons__orbisack_8h.php 
> <https://www.scipopt.org/doc/html/cons__orbisack_8h.php>
> https://www.scipopt.org/doc/html/cons__orbitope_8h.php 
> <https://www.scipopt.org/doc/html/cons__orbitope_8h.php>
> https://www.scipopt.org/doc/html/cons__symresack_8h.php 
> <https://www.scipopt.org/doc/html/cons__symresack_8h.php>
> All those are for binary variables.
> I occasionally encounter 0-1 ILP which have symmetries so large that 
> SCIP fails to (fully) detect them. I asked on the OR stack exchange 
> whether there is a standardized format for symmetry groups: 
> https://or.stackexchange.com/questions/13187/are-there-existing-formats-to-express-symmetries-between-variables-in-milp?noredirect=1#comment28418_13187 <https://or.stackexchange.com/questions/13187/are-there-existing-formats-to-express-symmetries-between-variables-in-milp?noredirect=1#comment28418_13187> and was pointed towards SCIP being potentially interesting for that.
> Is there some examples which use those constraints handlers?
> The documentation i linked above is to dense for me to make sense of.
> Sincerely,
> Johann-Tobias Schäg
> 
> _______________________________________________
> Scip mailing list
> Scip at zib.de
> https://listserv.zib.de/mailman/listinfo/scip