| This thesis addresses the optimization of large-scale scheduling problem, specifically in the area of refinery operations. Due to the high complexity of the entire system, the refinery scheduling problem is spatially decomposed into three subproblems: the crude-oil unloading and blending, the production-unit operations and the product blending and delivery. Comprehensive mathematical models are developed to describe the characteristics of each of those subproblems and solved efficiently due to the full utilization of time continuity.; An important consideration in scheduling and of particular interest to refinery operations is the issue of uncertainty introduced due to variability in product demands, prices, raw material availabilities, and process parameters. A rigorous methodology is thus developed that can be utilized to address the problem of uncertainty in decision-making process. The proposed approach is based on sensitivity analysis within a branch-and-bound solution methodology that results in the identification of the ranges of uncertain parameters where the schedules remain feasible and optimal, and the determination of a set of alternative schedules to cover the uncertainty space. In order to obtain schedules that are less sensitive to the model data, we developed a multiobjective robust optimization model to capture the trade-offs of economic optimality and decision robustness in the face of uncertainty and enable the generation of a number of alternative solutions.; These uncertainty analysis approaches are applied to the optimal operations of crude oil unloading and mixing, and gasoline blending and distribution to illustrate the importance of considering uncertainty in refinery scheduling operations. |
