Scheduling Crude Oil Operations in Refineries with Genetic Algorithm

With the hybrid characteristics of a refinery, it is very challenging to schedule crude oil operations. This work intends to solve this scheduling problem by decomposing it into two sub-problems hierarchically. At the upper level, a refining schedule is found, while a detailed schedule is obtained to realize it at the lower level. Given a refining schedule at the upper level, this work focuses on the detailed scheduling problem at the lower level. Based on a control-theoretic perspective, the problem is transferred to a problem of assigning charging tanks to distillers such that meta-heuristics can be applied. Then, a genetic algorithm (GA) approach is innovatively developed to solve it. An industrial case study is used to show the application of the proposed approach. It shows that the method works well and is applicable to real-life problems.