Research And Implementation Of Bus Intelligent Dispatching System Based On Data Mining

In recent years,with the intelligent development of public transportation,mass transit information has been accumulated in the bus operation and scheduling,these information have obvious big data characteristics,but they have not been fully explored,which leads some cities to use the experience or manual investigation to formulate the driving plan in the bus dispatching,which greatly reduces the bus operation efficiency.Therefore,it is of great value for the intelligent development of bus dispatching to explore the relevance and predictability through the collected mass transit information.Bus Rapid Transit(BRT),as the product of the intelligent development of the bus system,has advanced hardware equipment,fast transmission network and large data analysis application software system,which provides a relatively comprehensive data source for the research of the thesis.In view of this,the thesis takes the rapid transit BRT system as the research object to carry out the demand analysis and overall structure design,and researches and implements a bus intelligent dispatching system based on data mining.The main contents of this thesis are as follows:The thesis expounds the basic concepts of data mining overview,data mining architecture and related technologies required for system development,from the analysis of system functional requirements and non-functional requirements,the overall architecture,system architecture,system logic structure and system deployment scheme of the system are designed with BRT as the research object,which provides a theoretical basis for the system application.The thesis studies the data mining methods under the BRT intelligent scheduling system,through the comparative analysis of different collection methods of bus passenger flow,the passenger flow statistic instrument is selected as the passenger flow collection mode.Application of GPS systems for bus travel data collection.The thesis analyzes the passenger flow from the hourly passenger flow,the full-day passenger flow during the week,the seasonal and short-term passenger flow,the passenger flow in the upper and lower directions,and the uneven passenger flow in the section,according to the principle of BP neural network and the running process,the BP neural network traffic flow prediction model is established,and the prediction results are verified by MATLAB software examples.The thesis applies the SVR support vector machine to establish the bus travel time prediction model,and analyzes the prediction results through the SVR toolbox example.According to the passenger flow forecast and the travel time prediction,the BRT bus intelligent dispatch optimization model is designed to realize the best shift schedule of BRT lines.In order to improve the efficiency of system development,the thesis designed the BRT intelligent scheduling system by applying the JeePlus architecture.The system development platform is built,and the Sqoop and Flume technologies in the data mining platform are applied to complete the BRT basic information real-time collection and historical data transfer scheme.According to the analysis of system requirements,the main functions of the system such as driving planning,real-time monitoring,dynamic scheduling,data statistics management and data analysis are realized,and the unified access interface is designed to improve the linkage of the system database.Part of the function module of the system debugging and performance,verify system stability and security,debugging basic research to meet the design requirements,can provide reference for the the development of intelligent bus dispatch system in China.