Learning Objective Boundaries for Constraint Optimization Problems