Based on the above analysis, and taking into account the results of Lemma 3 which call for searching over a limited range of the m and n_j values for given values of k and s, we next present in an algorithmic form an efficient search procedure to determine the global optimal values for the different problem decision variables.