My datasets consist of large SAT instances. I am trying to find out how SCIP's default settings selects the branching rule to use.
There is a dataset that I have which consists of 7295 variables and 409834 constraints. After setting the time limit for 8h, using default settings and writing the statistics into a file, I see that SCIP has called inference branching.
However, when I tried this again with a smaller data set of 1040 variables and 3668 constraints, SCIP called relpscost branching instead.
Is there some documentation or code to what criterion SCIP uses to select the branching rule? I am currently unable to find it if that exists.
It is explained in detail in the documentation of branching rules, so I will just post a link here (section priority):