Data-driven Dynamic Path Optimization And Parking Guidance Models And Algorithms

Traffic congestion and parking are two major problems for travelers.Due to the influence of various factors,travel time and available parking spaces are highly dynamic and random.It is necessary to fully consider and effectively deal with the dynamics and randomness of the traffic network and recommend reliable paths and available parking lots for travelers,which is a subject that is worthy of further study.Aiming at the dynamics and randomness of travel time and available parking spaces in the traffic network,this dissertation first studies the distribution of travel time and the number of available parking spaces based on actual data,then establishes the models for finding reliable path and available parking space,and finally proposes the solution algorithms.Furthermore,the proposed models and algorithns are applied to real road network to find reliable path and available parking space.The main research contents and innovations of this dissertation are summarized as follows:(1)The extended shifted lognormal(ESLN)distribution is proposed to describe the grouped travel time.The model based on ESLN distribution and the solution algorithm based on the boundary of reliability are proposed to find the most reliable path with a given expected travel time.According to the analysis of real data,travel time is highly uncertain due to the various factors,such as holidays,day-of-week,time-of-day,and trafiic states,etc.,which often reduce the cumulative probability of travel time even in the same facility type(the same lane number and the same divided type).Thus,an aggregate approach is proposed to classify travel time data based on these influence factors.The distribution with the new aggregate approach is defined as the extended shifted lognormal(ESLN)distribution.Then,the link travel time model and the path travel time model are established based on ESLN distribution.A model based on ESLN distribution and the solution algorithm based on the boundary of reliability are proposed to find the most reliable path with a given expected travel time.Finally,the model and the solution algorithm are used to find the most reliable path in a real road network.The comparison between the ESLN distribution and the shifted lognormal(SLN)distribution shows the effectiveness and improvement of the ESLN distribution in finding the most reliable path.(2)The models based on ESLN distribution and the solution algorithms based on the boundary of travel time are proposed to find the latest departure time and the earliest arrival time with a given reliability.The reliable path problems with a given reliability can be divided into two categories:The Forward problem and the Backward problem.While the reliability and the departure time are given,the model for calculating the earliest arrival time based on the ESLN distribution is called the Forward problem.While the reliability and the arrival time are given,the model for calculating the latest departure time based on the ESLN distribution is called the Backward problem.Two mathematical programming models based on the ESLN distribution are proposed to describe the Forward problem and the Backward problem,respectively.Two corresponding solution algorithms based on the boundary of travel time are proposed to solve these two models.Case studies with a real road network in Beijing are given to illustrate the effectiveness of the proposed models and algorithms.(3)A model is proposed to predict the number of available parking spaces.And the parking guidance model considering the most reliable path and the algorithm based on the boundary of reliability are established.The distribution of the number of available parking spaces is analyzed based on historical data.It is found that the number of available parking spaces in the same hour and the same week are highly similar,but are quite different in different holidays,weeks and hours.Thus,a grouping method of available parking spaces is proposed based on holidays,weeks and hours.KS test shows that the number of the grouped available parking spaces obeys normal distribution.Then,a model based on normal distribution is proposed to predict the number of available parking spaces.The parking guidance model considering the most reliable path and the solution algorithm based on the boundary of reliability are established to find the most reliable parking lot and the most reliable path with a given departure time and a given expected travel time.Case studies with a real road network in Beijing are given to illustrate the effectiveness of the proposed model and algorithm.(4)The parking guidance models considering the earliest arrival time and the latest departure time are established.And the solution algorithms based on travel time boundary are proposed.Mathematical programming models and the solving algorithms based on travel time boundary are established for two kinds of parking guidance problems.1)To find the earliest arrival time and the corresponding reliable path with a given departure time and a given path reliability,a parking guidance model considering the earliest arrival time and a algorithm based on the travel time boundary are proposed.2)To find the latest departure time and the corresponding reliable path with a given arrival time and a given path reliability,a parking guidance model considering the latest departure time and a algorithm based on the travel time boundary are proposed.Case studies with a real road network in Beijing are given to illustrate the effectiveness of the proposed models and algorithms.(5)Reliable path and parking guidance system is developed.Reliable path and parking guidance system realizes the following functions under two scenarios:no parking demand and parking demand.When there is no parking demand,the system only calculates the reliable path with three eases:1)the most reliable path with a given travel time,2)the earliest arrival time with a given reliability,and 3)the latest departure time with a given reliability.When there is parking demand,the system can calculate the parking lot and the reliable path with three cases:1)parking guidance considering the most reliable path,2)parking guidance considering the earliest arrival tune,and 3)parking guidance considering the latest departure time.