Specifying branching priority in JuMP/CPLEX

[Claudio Contardo]

Hello,

As of release 0.14 of JuMP, it is still not possible to specify a branching priority for a MIP. I am trying to solve this single-source capacitated facility location problem by B&C using JuMP/CPLEX. In this problem, the location decisions usually have a much larger impact on the bound than the assignment decisions, so branching on those variables first is only natural. In this topic someone suggests to use the functions provided in the CPLEX module directly, however I do not know what is the right way to do so. A sketch of my code is something like:

model= Model(solver = CplexSolver())
@variable(model,y[1:m],Bin)
@variable(model, x[1:m,1:n],Bin)

############################
#add a bunch of constraints#
############################

####################################
#specify branching priorities, how?#
####################################

solve(model)

[Joey Huchette]

This is still a CPLEX-specific feature (as mentioned in the other thread). You’ll need to work with the CPLEX-level model object (internalmodel(model).solver) and the “linear indices” of the variables (accessible as, e.g. linearindex(y[1])).

-Joey

​

--
You received this message because you are subscribed to the Google Groups "julia-opt" group.
To unsubscribe from this group and stop receiving emails from it, send an email to julia-opt+unsubscribe@googlegroups.com.
Visit this group at https://groups.google.com/group/julia-opt.
For more options, visit https://groups.google.com/d/optout.

[Claudio Contardo]

Thanks Joey,

I was indeed capable of making it work :-)

Best regards,

Claudio

To unsubscribe from this group and stop receiving emails from it, send an email to julia-opt+...@googlegroups.com.