Research On Charging Station Recommendation Technology Based On Link Prediction For Cruising Electric Taxis

With the increasingly serious environmental damage caused by fossil fuels and the rapid development of electric vehicles,more and more electric vehicles have emerged.In order to prevent the electric vehicles from being exhausted on roads,an electric vehicle typically navigates to a charging station when its residual battery level is lower than a battery threshold,and the movement deviation from the destination(referred to as extra movement)is expected to be minimized as much as possible.As a special case of electric vehicles,electric taxis(ET)pick up passengers and the destinations are generally determined by the carried passengers.While an ET is cruising,its destination is always equivocal,and the minimization of extra movements from the potential destinations of cruising ETs becomes a vital issue.Besides,the movement of a cruising electric taxi is quite sophisticated,and it typically depends on the driver’s subconscious movement tendencies,driving habits,and historical passenger locations.Thus,the future destination is hard to be obtained,and this thesis exploits the links between ETs and charging stations through exploiting the historical trajectories.To this end,we propose a Charging station Recommendation Algorithm based on Link Prediction(CRA-LP)to recommend the proper charging stations for cruising ETs.In CRA-LP,the Adamic-Adar index with a temporal decay coefficient is designed to measure the similarities between ETs and charging stations.In addition,the historical trajectories of other ETs encountered by a cruising ET can be used for the recommendation of charging stations as well.Simulation results show that CRA-LP achieves preferable results in terms of some metrics,such as Area Under Curve and Average Extra Movement,which also indicates that CRA-LP can predict the future links between cruising ETs and charging stations accurately,and then recommend the proper charging stations to the cruising ETs.Based on the above work,this thesis also designs and implements a prototype of charging station recommendation.The system relies on Wechat Applet,implements the CRA-LP algorithm,and visually shows the recommended charging stations on the electric map.Some running results show that the prototype can easily recommend the proper charging stations to users,and display the recommendations on the map,with a small overhead and a stable execution.In this thesis,a link prediction algorithm based on the Adamic-Adar index with a temporal decay coefficient is used to measure the similarity between ETs and the charging stations,to recommend the proper charging stations.In future,more simulations will be conducted on some real datasets,and the proposed algorithm will be further improved in terms of the optimization of parameter settings and the algorithm efficiency.