Improved Bayesian Best-Worst Networks With Geographic Information System for Electric Vehicle Charging Station Selection

Abstract

Electric vehicle charging stations (EVCSs) are essential for solving the energy consumption and endurance anxiety problems of car owners. EVCSs also promote sustainable development in urban economies without relying on fossil fuels. This research proposes a hybrid approach that integrates the Bayesian network with best-worst method (BN-BWM) and a geographical information system (GIS) to address the site selection problem for electric vehicles (EVs). BN-BWM is employed to address the indicator system, which consists of nine criteria from three aspects. BN-BWM calculates the final distribution of the total preference of all decision-makers. Then, a GIS is utilized for spatial analysis and superposition analysis to determine appropriate sites for charging stations (CSs). The novelty of this study lies in the development of a new decision-making method based on the combination of BN-BWM and a GIS. This method is not only more innovative but also highly operational and convincing regarding the accuracy of the weight results. This research provides feasible and reliable ideas for the site selection and construction of CSs. It can also help EV companies and government personnel carry out strategic planning. The study verified the applicability and effectiveness of the developed hybrid method in sixteen administrative regions in Beijing. According to the results, 1) an indicator system consisting of nine criteria is established, and roads, charging stations, and slopes are identified as the most sensitive criteria for site selection; 2) Three alternative stations (ASs) are identified as the most suitable sites for the establishment of CSs.