|M.Sc Student||Shtern Shimrit|
|Subject||Robust Multi Echelon Inventory Control|
|Department||Department of Industrial Engineering and Management||Supervisors||Professor Boaz Golany|
|Professor Emeritus Aharon Ben-Tal|
|Full Thesis text|
Market uncertainty, manifested as demand fluctuations, causes the so-called “Bullwhip" effect, which makes supply chains unstable and costly. Market uncertainty can often cause such fluctuations in real world supply chains, but protecting them against the possible impacts is complicated since demands are unknown and can be estimated only roughly and inaccurately. We suggest a robust control model, which employs the General Robust Counterpart (GRC) framework developed by Ben-Tal and Nemirovski (2005) and apply it to a multi-echelon serial supply chain model described by Love (1979). The GRC uses implicit representation of uncertainty and min-max approach, which is considered rather conservative. The GRC is based on previous robust optimization methodologies, such as the Affine Adjustable Robust Counterpart (AARC) which was applied successfully, by Ben-Tal et al. (2005), to a single echelon inventory system. We show that our model yields lower cost and a more stabilized system than other models which do not explicitly address uncertainties. We also explore the characteristics of the GRC method and how to use these characteristics to obtain better results.