A novel distributed parallel algorithm for crude oil scheduling is presented in order to shorten the long computation time in the scheduling of a long scheduling horizon problem.The event-tree based modeling and optimization algorithms are adopted and the whole problem is divided into two sub problems (the portside sub problem and the refinery sub problem).Thus

the algorithm is proposed by analyzing the solving sequence of the two sub problems.The algorithm is based on master-slave distributed model.The master node solves the refinery sub problem firstly and sends the quality changeover sequence that is needed to start the portside scheduling to slave nodes

then the slave nodes send the results back to the master node.By comprehensive comparison of the results of the two sub problems

an optimized scheduling scheme will be obtained.The numerical example demonstrates that the algorithm is efficient in reducing the time overhead

especially when the crude oil distillation units(CDUs) have different feed quality requirements and the scheduling horizon is very long.For example when the horizon is 4 weeks

the sequential algorithm can not get a feasible solution in 48 h while the proposed algorithm can get optimal solution within 25 h.