Research On Differentiated Chargingstrategy Of Expressway Based On Usertravel Feature Classification

At present,China’s highway network is becoming more and more perfect,highway travel demand is increasing.There are obvious differences in travel demand of highway sections,which are mainly manifested in the fact that some sections are prone to congestion during peak hours,while few vehicles travel during peak hours.The unbalanced spatial and temporal distribution of traffic volume has led to the waste of resources in the road network.In order to solve the above problems,based on the analysis of massive toll data of highway network,this paper classifies highway users according to the differences in travel characteristics of highway users,and formulates targeted differentiated charging strategies according to the travel characteristics of each type of users.The price mechanism is used to adjust the travel time and path of users,so as to realize the balanced use of spatio-temporal resources of highway network and promote the maximum traffic efficiency of highway network.The main research contents and results include :(1)Highway network toll data preprocessing.The redundant data in the original toll data and the wrong data that do not conform to the traffic logic are identified.According to the characteristics of abnormal data,different preprocessing methods such as partial deletion,total deletion and application of web crawler technology are proposed to repair and process data,so as to ensure the true and effective toll data.(2)Construction of highway user travel characteristics index system.Based on the preprocessed toll data of highway network,according to the differences in travel characteristics of different users,seven travel characteristic indexes are constructed from the four dimensions of vehicle category,travel intensity,travel time and travel space,including vehicle category,travel times,weekend travel times,the first travel start period,the last travel end period,trajectory repetition rate and travel distance,and the extraction process and method of each index are formulated.Taking the toll data of highway network in Shandong Province as an example,the travel characteristics of its users are systematically described.(3)Construction of highway user classification model.According to the travel characteristics of highway users,K-means,fuzzy C-means and SOM algorithms are selected to classify highway users,and the clustering effect is evaluated by using the sum of squares of error,Silhouette Coefficient and Davies-Bouldin Index.Using the SOM algorithm with the best clustering effect,the expressway travel users in Shandong Province are divided into six categories: high frequency short distance travel users,intermediate frequency time stable travel users,intermediate frequency stable travel users,low frequency morning travel users,low frequency night travel users,low frequency long distance travel users,and the typical travel rules and characteristic parameters of each user are summarized.(4)The differential toll strategy of expressway is proposed.Based on the summary and analysis of the current differential toll strategy and implementation status of highways,combined with the travel characteristics and classification of highway users in Shandong Province,the differential toll measures of highways based on the travel characteristics of users are proposed.The road saturation and the average single-cycle benefit are selected as the evaluation indexes,and the simulation verification is carried out with an example.The research results will help to adjust the balanced distribution of traffic flow on expressway network,which is of great significance to improve the utilization efficiency of road network resources,and can provide reference for the implementation of fine differential toll policy based on user classification.