In order to alleviate the current situation of urban traffic congestion, as a supplement to conventional bus routes, microcirculation bus has more room for development. The purpose of this paper is to make full use of the information obtained from the large data of buses and taxis, accurately lay out the microcirculation bus lines, at the same time improve the bus coverage, reduce passenger travel time, and ultimately realize the efficient service of hierarchical public transport system. In this paper, Oracle software is used to process Xi'an bus IC card, GPS data and taxi GPS data. On the one hand, the passenger flow at bus stations is obtained by conventional bus data, and the operation status of bus lines and spatial accessibility of conventional buses are analyzed. The distribution area of microcirculation buses is determined by combining the nature of land use. On the other hand, the taxi GPS data is used to accurately acquire the passenger flow at bus stations. OD clustering of passengers' travel in this region provides theoretical support for setting up waiting stations on microcirculation bus lines. Based on the above analysis, taking the minimum travel cost of residents as the main objective, the bus company uses the minimum operating cost to transport the maximum passenger volume as the secondary objective to construct a bilevel programming model. Finally, the genetic algorithm is used to solve the model. The research results can provide a basis for the traffic management department to develop micro-circulation traffic lines. Key words: Microcirculation Bus; IC Card and GPS Data; Spatial Cluster Analysis; Double layer model.

CICTP