> At work i encounter a lot of "SOS #39; type problems - mostly when > i want to model "choose exactly one member of a group". Currently I > model this with binary variables, whose sum i constraint to be less > than 1. Modeling this using SOS seems like a better approach- more > compact, and also could yield better MIP results. > Are SOS #39;s supported in glpk?
SOS1/SOS2 constraints are not supported in glpk directly. However, you can model them using binary variables. See, for example: http://lists.gnu.org/archive/html/help-glpk/2007-06/msg00005.html > If not, any reason why? Constraints like SOS1, SOS2 (as well as alldifferent, for example) are peculiar to constraint programming, not to MIP. _______________________________________________ Help-glpk mailing list [email protected] http://lists.gnu.org/mailman/listinfo/help-glpk
