Worst-case analysis of the full load policy in the single link problem

The full load policy is a popular heuristic solution of the problem of shipping several products from a common origin to a common destination with the aim of minimizing the sum of inventory cost and transportation cost over an infinite time horizon. We show that, although this policy gives in the worst case an infinitely large error with respect to the optimum, there exists a class of instances in which it is better than the best single frequency policy. Moreover, we show that there exists an instance in which this policy allows to reduce significantly the total cost with respect to the best possible frequency-based policy with any number of transportation frequencies.