Abstract: The optimization of crude oil scheduling for coastal refineries was studied,and a mixed-integer nonlinear programming model based on discrete time representation was developed. An efficient evolutionary multi-objective optimization algorithm,non-dominated sorting genetic algorithm,was used to solve the model. Aiming at the high-dimensional decision variables in crude oil scheduling optimization problem, the periodic rolling optimization strategy was used to reduce the dimension effectively, and Deb criterion was used to deal with the problem effectively.The results showed that the optimized scheduling plans were all feasible, the minimum number of formula switching was stable at 12-15, and the minimum total berthing time of oil tanker was stable at 123-132 h,which showed that the model and the optimized algorithm could effectively deal with the crude oil scheduling problem of coastal refineries.It can provide an optimal method for crude oil scheduling optimization in refineries which depends on artificial experience.

Key words: crude oil scheduling, coastal refinery, multi-objective optimization, evolutionary algorithm