The Study Of The Travel Time Reliability

At present,the number of cars in Shenzhen has reached 3.22 million,and the road density is among the highest in the country,and urban road traffic is now oversaturated.For urban traffic roads,congestion is the result of the imbalance between traffic supply and traffic demand.This leads to the road can't afford to any interference.Even a little disturbance is likely to bring great inconvenience.And this instability also exacerbate the doubts of the traveler,unable to accurately determine their departure time.Therefore,the study of the reliability of travel time is very necessary to improve the travel satisfaction of travelers.In this paper,based on the license plate data was studied by analysis of urban road travel time reliability problems.It was based on link travel time distribution form to build link travel time the cumulative distribution function,and present a travel time reliability index.On this basis,the path travel time with certain positive correlation was calculated,and the cumulative distribution function of path travel time was given.The specific work mainly included the following aspects:First,the data was preprocessed in batches to eliminate the data noise.This research was based on the acquired license plate data,and a batch processing method was adopted to deal with the small batch travel time data.In batch data,we found the reasonable lower limit value,and eliminated reasonable outliers' time,and determined the need to keep the numerical.This effectively improved the data demising threshold methods,and effectively removed the noise of the data,then improved the accuracy of data.Secondly,Monte Carlo simulation algorithm was used to study the reliability of road travel time.Through the analysis of link travel time distribution,we found that the road travel time was bimodal distribution,and in different sections and periods of time,travel time distribution form had a certain difference.Through the data sample of this study we found that travel time was not subject to fall and logarithmic fall.Because of the complexity of the data,it was difficult to solve the travel time reliability by mathematical analytical method.Monte Carlo simulation algorithm can be used to solve the problem of reliability calculation under the characteristics of specific analytic functions.Thirdly,the study on the correlation of travel time of road sections was conducted.It was found that adjacent road section travel time had a certain positive correlation,and not all of the routes were independent identically distributed.Besides,through the method of regression analysis to the existence of correlation was verified.Finally,the model to solve the travel time was built.On the premise of a certain correlation between the travel time data of adjacent road sections,the travel time correlation matrix of adjacent road sections was constructed to describe the relationship of travel time between adjacent sections.Based on Monte Carlo simulation algorithm,the path travel time distribution function and cumulative distribution function were constructed,and the corresponding route travel time of different reliability was given to provide reliable travel information for travelers.