(555f) A Novel Solution Strategy with the Event-Tree-Based Modeling Approach for Crude Oil Scheduling
AIChE Annual Meeting
Wednesday, November 10, 2010 - 4:55pm to 5:15pm
Crude oil scheduling is an important sector for the supply chain sustainability. In order to quickly obtain an optimal scheme for the large-scale crude oil scheduling, a novel solution strategy is proposed with the event-tree-based modeling approach. The new modeling approach is dynamically solving the event-tree induced real-time models consisting of all feasible states in the scheduling horizon, which is completely different from mathematical programming techniques. The event triggered rules are used to search for the single node of the event-tree model, where the continuous time presentation is adopted to divide the scheduling horizon. To reduce the complexity of the problem, the entire scheduling problem is decomposed to an unloading sub-problem and a charging sub-problem. Finally, the depth-first search algorithm (DFSA) is employed to search the optimal solution. The major improvement of this work is that it greatly improves the searching efficiency compared with the breadth-first search algorithm (BFSA). Case study results show that the event-tree-based approach with DFSA is an effective and efficient way for dealing with the crude oil scheduling problems.