Minimizing total tardiness on two uniform parallel machines considering a cost constraint.