I am running two MIP models using Gurobi, I call them A and B. B is the relaxed problem of A, and B has fewer variables, fewer constraints, fewer nonzeros, and fewer rows than A. Is that possible that solving B requires a higher RAM than solving A? (As I said B is the relaxed problem of A, and the two models are similar).
By the way, I run a set of numerical experiments. I found that solving the relaxed problem is much faster than solving A.
Please sign in to leave a comment.