The Optimization Of Public Transportation Sites And Route Based On Taxi GPS Data

With the steady growth of Chinese economy, the rapid development of urbanization and the expansion of urban scale, the demand of motor vehicle are also increase. But due to the limited urban space, relatively slowly construction of traffic infrastructure, continuously increasing of urban residential population and the urban transportation demand, urban traffic problem has become increasingly serious. Urban traffic problems, such as traffic congestion, uneven distribution of traffic resources, energy waste, have become one part of the key factors of the economic development and the promotion of social operation efficiency. Based on the investigation and analysis of the existing urban traffic situation, we have found that, bus and taxi which are main urban residents travel tools still exist large problems in the public bus-stop location and taxi route planning, such as the setting of bus stops cannot be good to meet the actual travel demands of urban residents, and the high empty loading rate of taxi cause serious waste of urban traffic resource, these phenomena lead to problems such as traffic congestion at the same time.In the era of big data, the taxi trajectory data based on the Global Positioning System(GPS) provides the possibility to solve traffic problems including bus-stop locations and taxi route recommendation through big data analysis. This study seeks the travel patterns of people and the pick-up characteristics of the driver through the analysis of taxi GPS data. And through optimizing clustering algorithm and prospect theory, we designed the bus stop of the city and recommended the driving path of taxi, so as to alleviate the existing traffic problems and effectively improve the urban traffic operation efficiency.Under the background of the increasingly serious urban traffic problem, we proposed the C-DBSCAN clustering algorithm, then designed the driving route of city bus and recommended the driving path of taxi so as to solve the traffic problem through the analysis of massive taxi GPS data. The main contributions of this paper are as follows:1. Proposed the C-DBSCAN clustering algorithm:This paper based on the CDBSCAN clustering algorithm, we introduced spherical distance to measure the distance between two data points(i.e., pick-up or drop-off locations) and calculated a cluster center by averaging the latitudes and longitudes of all the data points in a cluster.2. Determined bus-stop locations by using clustering algorithm: This paper utilizes the C-DBSCAN algorithm to draw the travel patterns of people and find out the pick-up area based on taxi GPS data, so as to discover candidate bus-stop locations.3. Recommended the driving path of taxi by using prospect theory: Through analyzing the GPS data of taxi drivers and the driving path, we gain the drivers’ different expected return, the pick-up probability and the return of driving path. Base on prospect theory, we used the utility of return which is set as the value function and the pick-up probability which is set as weighting function, so as to get prospect value of driving path in certain time slot for taxi drivers, and recommended the optimal utility driving path which is considered the drivers’ personalized utility(measures the drivers psychological satisfaction of the recommend results). Through the path recommendation, this study expect to achieve the optimal allocation of resources.In summary, this paper proposed the C-DBSCAN clustering algorithm, then designed the driving routes of city bus and recommended the driving paths of taxi so as to solve the traffic problem through the analysis of massive taxi GPS data. We believe that our research will provide some beneficial reference for solving more similar problems in future.