Optimal long-term prediction of Taiwan’s transport energy by convolutional neural network and wildebeest herd optimizer

Abstract

In this paper, an optimized model based on Convolutional neural network (CNN) models is proposed for transport energy forecasting in Republic of China (Taiwan). This study employs various self-reliant parameters including individual, gross domestic product, vehicle registrations’ number, value of passenger transport, and oil price for modeling. To provide an optimized result of CNN, it is combined with newly defined wildebeest herd optimization algorithm and the model is called WHO/CNN. The model results are compared with multiple regression model and the basic CNN to show its higher efficiency in prediction of the transport energy data. Simulations are compared based on MTOE as well as R2to indicate the superiority of the suggested model. Final results show that based on the prediction, however, the demand of transport energy in Taiwan will not increment excessive, being about 37.2 MTOE in 2020 while gross domestic product increasing at the time of the same duration is almost high. The predicted outcomes are in accordance to ROC’s green growth strategy, which needs CO2reduction gas as it is feasible. However, there is minor growing of demand prediction of the transport energy, ROC requires to continue the CO2emission decreasing.

Keywords

• Transport energy prediction
• Artificial neural network
• Wildebeest herd optimization algorithm
• Republic of China