Abstract: A novel modeling and optimization approach for crude oil scheduling based on event trees is presented in this paper. Based on the natural language modeling approach, a set of events and the state of the system at an individual time are defined. Event triggered rules with respect to each event are built up. Each node of the event tree stands for a state of the system at a certain time. The event tree is composed of all feasible nodes during the scheduling horizon. Event triggered rules and operating rules are used to restrict the search space. The crude oil scheduling problem is decomposed into a portside scheduling subproblem and a refinery scheduling subproblem. This method is successfully applied to a problem over a horizon of one week involving 2 vessels, 6 portside tanks, 6 refinery tanks, 2 CDUs, and 6 crude oil types.

Key words: crude oil scheduling, continuous time, event tree, natural language, modeling