Crude Oil-Import Portfolio Optimization Based on Interval Prediction Data

Abstract

The optimization of crude oil-import portfolio is a hot research issue in the field of energy security, which is related to the implementation of national strategy and development of economy. Forecasting the demand of crude oil-import is the basis for portfolio optimization. Therefore, this paper innovatively introduces the decomposition hybrid interval prediction method and proposes a multi-objective programming model so as to provide decision-making support for the formulation of crude oil-import portfolio scheme. Under the constraints of volume, price and risk, the minimum cost and risk of importing crude oil are achieved. Furthermore, by introducing optimization parameters and risk preference factors, and setting different scenarios for numerical simulation, the results show that (1) the decomposition hybrid prediction methods perform better in predicting the import demand of crude oil. Since the prediction accuracy of the decomposition hybrid prediction method is much higher than the single prediction model, in terms of MAPE and RMSE test; (2) NSGA-II algorithm can effectively solve the multi-objectives problem of crude oil-import portfolio.