A novel two-stage location model of charging station considering dynamic distribution of electric taxis

Abstract

Electric taxi (ET), with zero exhaust emission, plays a vital role in low-carbon sustainable development of urban public traffic to reduce fossil fuel consumption and air pollution. As the key part of ET development, charging station (CS) directly impacts ET’s widespread application by figuring out the charging difficulty of ET. Therefore, to tackle the location issues of electric taxi charging station (ETCS), a novel two-stage location model is built in consideration of ET dynamic distribution and charging requirements in this paper. Firstly, taking the complex features of ET into account, such as dynamic location and dispersed distribution, the dynamic distribution clustering model of ET is formed based on k-means method, by which lots of scattered ET locations are grouped into certain clustering blocks. Secondly, according to aforementioned blocks, the location model of ETCS is established with barycentric method to get the location of ETCS, in view of both the latitude and longitude location and charging demands of each clustering block. Finally, the two-stage location model is employed in the case study of ETCS location in Chengdu, China. The result shows that the chosen location meets the charging requirements and the characteristic of taxi distribution in Chengdu, proving the validity of the model. The two-stage location model is an improved approach of traditional static barycentric method by processing ET’s dynamic distribution and decentralized charging requirements with k-means clustering algorithm, and is of theoretical and practical significance to provide quantitative ETCS location solutions to help deal with actual CS location problems in urban management.