Flower pollination algorithm approach for the transportation energy demand estimation in Turkey: model development and application

Abstract

This study proposes a new optimization technique to estimate the Transportation Energy Demand (TED) employing the Flower Pollination Algorithm (FPA). The TED estimation models were developed based on three parameters, which are Annual Vehicle-Km (AVK), Gross Domestic Product per Capita (GDPperC), and Carbon-dioxide (CO2) emission according to the linear, power and quadratic forms. These three parameters were determined by the WEKA data mining software program among nine parameters. Randomly selected 80% of historical data for 47 years, from 1970 to 2016, were used for the training of the algorithm, and the remains were used in the testing stage of the models. The performances of the models were evaluated according to six different statistical criteria. Transportation energy demand forecasts by 2035 were carried out using three different scenarios using the TED estimation models. According to the scenarios, it is predicted that the transportation energy demand in Turkey will have doubled by 2035 in comparison with 2016. The FPA approach has been successfully applied in the development of the TED estimation models. The most important impact of this study is to help the creation of strategic action plans for energy policies in the transport sector and to contribute to the more efficient use of limited energy resources in the country.