Research On Decision Support Technology For Expressway Operation Using Data Mining

In view of the growing information for operation and management, network toll and monitoring system has accumulated a lot of toll data and vehicle records. However, the existing expressway operation and management technologys still stay in inquiring data and statistical analysis,these can not meet some intelligent management needs,such as road conditions forecastion, information services and vehicle audit. Based on expressway historical toll data and around the concept of intelligent expressway operating management platform,this paper discusses some ways that use mining data to discover a lot of laws of driving vehicles.On this basis, the paper focuses on the traffic forecast, travel time prediction and audit of evasion vehicle these three hot issues,thereby laying the foundation for intelligent decision support in expressway operation management. The paper’s main research and related innovations included:(1) Traffic flow prediction has been the focus of research in the transportation field,To resolve this issue is of great significance for traffic guidance in the toll station’s surrounding area.Taking the expressway macroscopic dynamic traffic flow model as prototype, This paper propose a traffic flow forecasting model based on BP neural network and achieves short-term traffic flow forecasting on expressway stations.(2) Paper designs a fuzzy travel time prediction model which merges a variety of factors based on average travel time under historical data. and uses to predict the single vehicle travel time.The model is different from the previous which emphasis on the collection of traffic flow data,taking the vehicle characteristics、driving time and other factors which affect the travel time into consider, and achieves neural network fuzzy travel time prediction.Model has a good prediction results and shows some practicality.(3) Paper designs a mining model by make use of the association between toll record characteristics and toll evasion phenomenon and achieves auditing analysis of those behavior oftoll evasion.It realized the design and application of the founded model by using mining tool Weka.The results of the model application shows a high audit performance for the changing category,changing car and other behavior of toll evasion.This paper focuses on mining data and analysising on specific issues,the mining results show some practicality and provide reference data for realizing traffic flow guidance and vehicle audit.Experimental results show that mining road vehicle information and laws with historical toll data has important value and significance for expressway operation’s decision support.