Robust scheduling for multi-product pipelines under demand uncertainty

Data uncertainty is an inherent component of many real-world problems, and ignoring the uncertainty makes the proposed programs impractical or inefficient. Robust optimization, regarding the data uncertainty, provides a program that remains feasible for almost all situations. This paper studies the problem of scheduling of an oil pipeline that connects a single refinery to a distribution center (DC) under demand uncertainty. In the first phase, a deterministic mixed integer linear programming (MILP) model is presented for pipeline scheduling and inventory management at a DC. Then, by means of the Γ-robustness approach, the deterministic model is extended to a robust formulation. The robust model is implemented on a real-world multi-product pipeline and, by varying the level of conservatism, trade-off between robustness and cost is analyzed. Finally, by generating the random samples for product demand, the average of surplus or shortage of inventory for each level of conservatism is estimated.