Research On Taxi Service Strategies Based On Large-scale GPS Trajectory Data

At present,in the field of urban transportation,GPS equipment has been widely used.For example,almost every taxi is equipped with GPS terminal equipment,which will transmit taxi’s position and operating status to the Traffic Management Department for real-time monitoring.In fact,swarm intelligence information such as service strategies of taxi drivers is hidden in the large-scale taxi GPS data,making the best use of this information,on the one hand,can induce taxi drivers to improve operation mode and increase income,on the other hand can help Traffic Management Department to improve the efficiency of taxi system.However,the current analysis methods of taxi GPS data are relatively primary;while the taxi GPS data is in large-scale,for example taxi GPS data of Xi’an City in one year is approximate TB level.The practice proves that the traditional information processing platforms have not been able to analyze the GPS data of the above-mentioned scale effectively.Aiming at the above-mentioned problems,this paper mines and analyses the taxi service strategies from GPS data based on the big data platform Hadoop.Hadoop is currently a popular big data platform,which can accomplish the analysis of data in TBs via parallel computing architecture,and because of the advantages of open source,scalability,low cost and easy programming,it has become a fact standard in the field of big data processing.Taxi service strategies are the mathematic model of taxi drivers’ service process,mainly refers to passenger-searching strategies,passenger-delivery strategies and service-region preference.After data preprocessing,this paper first extracts day shift driver’s personal trajectory,next quantifies driver’s income,then gets sample drivers with high income and middle income,after that identifies and calculates the service strategies according to the model,finally analyzes the relationship between service strategies and income.In the test case,this paper makes a detailed analysis and comparison of the similarities and differences of the passenger-searching strategies,passenger-delivery strategies and service-region preference of the sample drivers in different time slots.