In the new global economy, the inventory control has become a priority for the
supply chain management. Safety stock is the sole way to fight against the
demand and the supply uncertainty, so determining the amount of it that must
be kept along the network to holistically minimize the risk of disruption while
maximize the profit is a critical issue. For the supply side, the focus is held in
the lead time variability which can significantly vary depending on the part of the
supply chain or new inconvenient facts could relevantly affect the lead time.
Even so, developed models have forced to assume a certain value or a
distribution for the lead time, yet this is risky. Historical data is often unreliable,
not available or insufficiently representative. Therefore, a new model based on
the Guaranteed Service approach and combined with robust optimization
techniques is proposed, working with the lead time volatility without assuming
any specific distribution. Interesting features arise from the new model such as
the smooth tractability of the problem, the facile computational skills required or
the lack of resources needed. This approach has been formulated and tested,
as well as the Guaranteed Service when a distribution is assumed for the lead
time and the original model. Then, the performance of the three of them has
been compared in order to find the correct way to deal with uncertain lead time.
Finally, the Robust Guaranteed model benefits promise better security to
companies and it also provides a powerful tool to manage the risk from the
supply side.
