A remark on the computation of optimum media schedules