In linear programming, the optimum solution is achieved at a vertex except:
1) When the feasible region is unbounded
2) When the objective function is parallel to one of the constraints
3) When the feasible region is empty
4) When the objective function is linear