A Topological Study of the Upward Refinement Operators in ILP