On the Value of the Cost of Optimal Control Problems for Constrained Difference Inclusions