Loading...

Archive

    25 June 2022, Volume 22 Issue 3 Previous Issue    Next Issue

    For Selected: Toggle Thumbnails
    Platform Competition Strategy Based on Social Network Structure Characteristics of Online Ride-hailing Market
    SUN Qi-peng, QIAO Jia-lu , ZHANG Kai-qi , SUN Jia
    2022, 22(3): 1-6.  DOI: 10.16097/j.cnki.1009-6744.2022.03.001
    Abstract ( )   PDF (1528KB) ( )   PDF(English version) (651KB) ( 200 )  
    To study how the online ride-hailing platform selects effective strategies in the highly competitive industry to realize fair market competition, the paper investigations the competitive relationship between online ride-hailing platforms from the dual perspective of the social and econometric networks. Firstly, the competition network of the national online ride-hailing market is constructed, and the network topology is analyzed by selecting indicators such as degree centrality, structural hole, and core edge structure. Then take the practical revenue data of 112 online ridehailing platforms as an example, this paper integrates the network topology attribute into the econometric model and empirically analyzes the relationship between the network index of online ride-hailing competition and platform revenue. The results show that the competitive network of ride-hailing platforms in China is not balanced at present. The core platforms have dense network relations, where most of the platforms are dispersed in the network, and a few of them are on the edge of the market. And the platform's topological position in the network is closely related to the platform's revenue and the core-edge structure has the most significant effect on revenue. Finally, the strategies of using cutting-edge technology, establishing horizontal and hybrid alliances, and providing differentiated services are proposed to improve the market competition.
    References | Related Articles | Metrics
    Differential Pricing Optimization of High-speed Railways Based on Discrete Price
    WANG Yu , MIAO Lei
    2022, 22(3): 7-14.  DOI: 10.16097/j.cnki.1009-6744.2022.03.002
    Abstract ( )   PDF (1376KB) ( )  
    Different from most existing studies focusing on continuous pricing, this paper proposes a differentiated pricing optimization model for high-speed railways based on discrete prices. Firstly, a set of discrete discount rates is set, and railway passengers are classified according to their price sensitivity. Considering that the departure and arrival times in the different periods bring additional opportunity costs to passengers, a generalized cost function for each group of passengers is obtained, which is summarized as the overall passenger elastic demand function. We develop a linear integer programming model with multi trains and multi sections with an objective of revenue maximization. Finally, a PSO algorithm is used to solve the model. The numerical results show that the optimization model proposed in this paper can significantly improve the overall revenue. Moreover, the case using the continuous optimal price in most pricing theories can be regarded as a special case of the discrete pricing optimization problem. The discrete pricing fits the practice better, which should be attracted more attention
    References | Related Articles | Metrics
    Residents' Travel Mode Choice Behavior in Post-COVID-19 era Considering Preference Differences
    YANG Ya-zao, TANG Hao-dong, PENG yong
    2022, 22(3): 15-24.  DOI: 10.16097/j.cnki.1009-6744.2022.03.003
    Abstract ( )   PDF (2001KB) ( )   PDF(English version) (2509KB) ( 189 )  
    In order to explore the choice behavior of residents' travel mode in the post-COVID-19 era, a choice behavior experiment was conducted. A mixed Logit model and a latent class conditional Logit model of travel mode choice were constructed based on the data obtained from questionnaire surveys. Stata software was used to calibrate the model parameters, and the main factors influencing residents' travel mode choices were obtained. The results show that both models reflect the heterogeneity of individual travel mode choices. Compared with the mixed Logit model, the latent class conditional Logit model has an improvement of 13% in the goodness of fit and an increase of 3.03% in the prediction accuracy, which provides an effective tool for analyzing individual heterogeneity of travel behavior under public health emergencies. The latent class conditional Logit model divides residents into four and five groups according to the two scenarios of low and medium risk areas. From the perspective of travel mode attributes, the waiting time and the traveling time have become the most important influencing factors for residents to choose the travel modes. From the perspective of personal socio-economic attributes, women with higher incomes are more inclined to choose private cars to travel. The older are more sensitive to travel costs, and men are more willing to choose bus and subway travel.
    References | Related Articles | Metrics
    Impacts of Built Environment on Competition and Cooperation Relationship Between Taxi and Subway Considering Spatial Heterogeneity
    CHEN Qi-xiang , LV Bin , CHEN Xi-qun , HAO Bin-bin, HE Jia-xi
    2022, 22(3): 25-35.  DOI: 10.16097/j.cnki.1009-6744.2022.03.004
    Abstract ( )   PDF (4831KB) ( )  
    This study investigates the impact of built environment on the relationship between taxi and subway considering spatial heterogeneity. The study area is divided into grid cells and six subcategories of transportation system attributes and land use system attributes are selected as variables to describe the characteristics of urban built environment. The competition and cooperation relationship between taxi and subway are defined as SCPE (subwaycompeting), SE (subway-extending), and SC (subway-complementing) based on the spatial relationship of origin/ destination of taxi trips and subway stations. The multi-scale geographic weighted regression model (MGWR) is used to analyze the influence mechanism and spatial effects of built environment on the coopetition relationships (SC, SCPE, SE). The case study of Lanzhou city shows that the SC, SCPE and SE have significant spatial heterogeneity. The MGWR can characterize the spatial heterogeneity and scale difference between the coopetition relationship and variable dependencies and produce a more reliable estimate result. Moreover, the spatial impact of the built environment elements on the SCPE mode is relatively stable. The SC mode is closely related to bus stop density and road density, and there is high degree of spatial heterogeneity. The SE mode is also closely related to bus stop density.
    References | Related Articles | Metrics
    Operation Efficiency Evaluation on Urban Taxi Considering Energy Types
    CHEN Guang-hui , JI Hao , JIN Fu-qiang, GUO Qing-e , JIA Bin , SU Bing
    2022, 22(3): 36-44.  DOI: 10.16097/j.cnki.1009-6744.2022.03.005
    Abstract ( )   PDF (1536KB) ( )  
    With an analysis of expected operating input and non-expected operating input, expected output, and nonexpected output of fuel gas, electric, and methanol taxis in Xi'an, an evaluation model on operation efficiency is established by DEA- Malmquist considering energy types. The GPS trajectory data of 1000 methanol taxis, 1000 electric taxis, and 1000 gas taxis, which were randomly selected from Xi'an Taxi Management Office in 14946 taxis from September to December 2020, were taken. The travel demand and characteristics of passengers on workdays and weekends are found from processing GPS trajectory data, and the difference in the operation efficiency between the three types of taxis is analyzed. The results yield guidance for government decision-makers to choose different energy types of taxi operation decisions
    References | Related Articles | Metrics
    Dynamic Fleet Management of Shared Autonomous Vehicles with Rolling Horizon Optimization
    CHEN Yao , BAI Yun , ZHANG An-ying , MAO Bao-hua , CHEN Shao-kuan
    2022, 22(3): 45-52.  DOI: 10.16097/j.cnki.1009-6744.2022.03.006
    Abstract ( )   PDF (1718KB) ( )  
    The shared autonomous vehicle (SAV) is an essential component in future urban transportation systems. This paper investigates an optimization approach to the dynamic operationof a SAV fleet with stochastic demand. A timespace network is first constructed to characterize the fleet management problem. Different types of time-space arcs are generated to indicate the vehicle-trip assignment and empty vehicle relocation. Under the framework of approximated dynamic programming, this paper develops a mathematic programming model to maximize the operational profit, in which the flow of nodes is taken as vehicle states and the flow of arcs is taken as decision variables. The rolling horizon optimization, also referred as lookahead policy, is designed for the optimization problem. A stochastic program with a lookahead horizon is developed and solved by the CPLEX solver. A numerical case study is performed with the Sioux Falls network. The rolling horizon optimization approach can provide effective operational decisions of dynamic fleet management. Considering the computational time limit, a long lookahead horizon with a medium- size sample would produce better optimization results. The objective of maximizing the operational benefit while minimizing the passenger waiting time would also result in more effective decisions of the dynamic fleet management.
    References | Related Articles | Metrics
    Identification of Freeway Ramp Influence Areas Based on Multi-point Loop Data
    LI Yan, ZENG Ming-zhe, ZHU Cai-hua, WANG Fan , DENG Ya-juan
    2022, 22(3): 53-62.  DOI: 10.16097/j.cnki.1009-6744.2022.03.007
    Abstract ( )   PDF (3528KB) ( )  
    To accurately identify the influence of the freeway ramp on mainline traffic, a quantitative method is proposed based on the speed fluctuation characteristics. The improved weighted speed permutation entropy (IWSPE) is established to quantify the influence of freeway ramp under various levels of services (LOSs). The spectral clustering method is selected to determine the impact threshold of the freeway ramp using the proposed indicators. Using the data from 99 loop detectors located at parallel merge ramp and direct diverge ramp in G5 and G55 freeway of China, the results show that the proposed method can identify the influence brought by the ramp. The influence of the merge ramp to the outermost lane is 4%~69% higher than it to the secondary outer lane. The diverge ramp has an influence of 6%~ 29% higher to the outermost lane than the secondary outer lane under LOS A to C, while this influence is 10%~13% lower under LOS D to F. The influence area of the merge ramp ranges from 350 m upstream to 550 m downstream, with the core influence areas of 160 m upstream to 100 m downstream and 180~270 m downstream. The influence of diverge ramp to mainline traffic ranges from diverging point to 850 m upstream, with the core influence areas of 750~ 850 m, 450~600 m, and 100~300 m. The proposed method can contribute to the design, management and control, and reliability of traffic simulation of freeway ramp areas, which can effectively reduce the impact of the ramp.
    References | Related Articles | Metrics
    Route Choice Behavior of Urban Rail Transit Passengers Based on Guidance Information
    XU Xin-yue , XIE Lan-shi-yu
    2022, 22(3): 63-73.  DOI: 10.16097/j.cnki.1009-6744.2022.03.008
    Abstract ( )   PDF (1950KB) ( )  
    To explore how to provide guidance information for urban rail transit passengers, this paper considered the interaction between path attributes with the time and ways of information provision and update rate, and it proposed a framework for passenger route choice behavior based on guidance information. An improved mixed utility-regret model was established, in which the differences in the wishes of passengers to follow the mixed decision rules and the perception of each attribute are taken into account. The route attributes and guidance information are both incorporated into the scenario design, and the key parameters in the model were estimated based on passengers' route choice behavior survey. The results show that the mixed decision rule has resulted in a high fit degree (where the adjusted goodness ratio reached 0.396). The analysis on passengers' preferences for information provision indicates that, compared to downloading the mobile APP, passengers prefer to receive guidance information through social media. By introducing the time value of the information update rate, it is shown that female, middle- aged, and non- commuter passengers are more likely to accept higher frequency guidance information services. Based on the elastic value of each attribute, it indicates that the way of providing information and the update rate of information should be used as an auxiliary means for managers, and the information provision can play a greater role in specific scenarios (such as peak congestion, emergencies). These research results reveal how the way of guidance information provision affects the route choice of passengers, which is helpful to design accurate and efficient guidance information strategies for the organization and management of passenger flow.
    References | Related Articles | Metrics
    Electric Vehicle Charging Induction with Minimization of Negative Effects
    GE Xian-long, LI Ting , WANG Bo , YIN Zuo-fa
    2022, 22(3): 74-83.  DOI: 10.16097/j.cnki.1009-6744.2022.03.009
    Abstract ( )   PDF (1846KB) ( )  
    With the implementation of environmental protection policies such as "carbon peak" by the government, electric vehicles (EV) have developed rapidly with the advantages of energy saving and environmental friendliness. Due to the short cruising range, long charging time of electric vehicles and the space-time mismatch between recharge demand and charging pile in the road network, a series of negative effects have been found such as long queuing time and drivers' anxiety on the mileage ranges. This paper introduces incentives to achieve the optimal energy supplementation scheme and proposes a bi-level optimization model for electric vehicle energy supplementation to minimize the negative effect of EV charging on the road network. The upper layer is the induced excitation model which minimizes the negative effect of road network charging. The lower layer is a mixed road network equalization model with charging station selection. The genetic algorithm is designed to solve the upper model and the Frank-Wolfe algorithm is used to solve the lower model to obtain the optimal induction scheme of the energy-replenishing vehicles in the road network. The classic Nguyen-Dupius road network is taken as an example to verify and conduct sensitivity analysis. The results show that although the proposed charging induction model increases the incentive cost of planners, the total social charging negative effect cost is reduced, which proves the effectiveness of this model.
    References | Related Articles | Metrics
    Emergency Vehicle Dynamic Path Selection Based on Two-stage Objective Optimization Model
    YANG Feng , CHONG Da-shuang
    2022, 22(3): 84-92.  DOI: 10.16097/j.cnki.1009-6744.2022.03.010
    Abstract ( )   PDF (1787KB) ( )  
    This paper investigates the emergency vehicle routing problem when the traffic network changes dynamically and proposes a two-stage scheduling optimization model for emergency vehicle routing. The two-stage optimization model with the maximum path reliability and the shortest travel time is developed by combining the dynamic situation of the road network and the characteristics of emergency rescue. The initial population of Cuckoo algorithm is improved by chaotic search and the frog leaping algorithm is added to improve the local search operation. The hybrid Cuckoo algorithm is designed to improve the global optimization ability. Taking a regional road network in a city as an example, the real-time traffic data of the regional road network and the data obtained from field driving are applied to the model and solution algorithm. The results show that the dynamic reliable path from the starting point to the rescue point can be successfully obtained using the proposed model and algorithm. For the same driving path, the maximum error of the shortest travel time solved by the proposed method and the shortest travel time obtained by field driving is no more than 8%, indicating good feasibility of the method. Compared to the particle swarm optimization algorithm and the classical Cuckoo algorithm to solve the K shortest path, the hybrid Cuckoo algorithm can obtain the shortest travel time with fastest calculation speed, which shows that the hybrid Cuckoo algorithm has the best solution and performance.
    References | Related Articles | Metrics
    Customized Bus Route Optimization with Price Window
    YUE Hao , LIU Jian-ye, LI Chong-nan
    2022, 22(3): 93-103.  DOI: 10.16097/j.cnki.1009-6744.2022.03.011
    Abstract ( )   PDF (2109KB) ( )  
    Considering the welfare of public transportation and the trade-off between passengers and bus enterprises, the customized bus route optimization is studied under a floating pricing policy to ensure the operating cost of the enterprise and serve more passengers. Firstly, the concept of price window is introduced to describe passengers' willingness to pay, and the characteristics of travel demand based on temporal-spatial window and price window are defined. An assumption rule of over-equalized fare payment is proposed so that passengers pay the least cost on the basis of guaranteed travel. Secondly, an integer linear programming model is established with a price window, in which the temporal- spatial window and price window of passenger travels are set as input variables. The objectives are to minimize the operating cost and to maximize the number of passengers served. The operating cost of the enterprise is guaranteed by imposing the operating profit constraints. Two sub-problems are solved in an integrated manner, namely, passenger-to-vehicle assignment and vehicle routing. Finally, GAMS software is used for modeling and solving, and an example is analyzed on the Sioux Falls network. The results show that the introduction of the floating pricing policy can serve more passengers, and increase the operating income of the enterprise. The number of passengers served and the operating income will increase as the upper bound of the price window increases to a certain level.
    References | Related Articles | Metrics
    Feeder Bus Routes and Frequency Optimization Based on Mixed Integer Nonlinear Programming
    SONG Li-jing , BAI Tong-zhou , HE Yu-long , CHEN Yan-yan, LIU Xue-jie , MA Teng-teng
    2022, 22(3): 104-111.  DOI: 10.16097/j.cnki.1009-6744.2022.03.012
    Abstract ( )   PDF (1527KB) ( )  
    The optimization of feeder bus in terms of network design and frequency setting is an important topic about the coordinated development of rail transit and ground transit. This paper analyzes the model construction, planning methods, and model algorithms deficiencies in previous studies. The study develops an optimization model for feeder bus operations based on the mixed integer nonlinear programming, and the model is solved by reconstructions. Then, the model and algorithm are verified through case analysis. The results showed that: The proposed model in this study includes a multiple to multiple (M to M) constraint, which is closer to the real transit operation conditions. Compared to the Depth-first search (DFS) algorithm using the same topology road network, the proposed model reduces the calculation complexity and is feasible and effective in optimizing the feeder bus routes and departure frequencies
    References | Related Articles | Metrics
    Train Capacity Allocation Strategy and Optimization Model for an Oversaturated Metro Line
    SHI Jun-gang , QIN Tan , LI Xiang , YANG Li-xing, YANG Xiao-guang
    2022, 22(3): 112-119.  DOI: 10.16097/j.cnki.1009-6744.2022.03.013
    Abstract ( )   PDF (2419KB) ( )   PDF(English version) (559KB) ( 117 )  
    To alleviate the extreme congestions of local stations in oversaturated metro lines, this paper proposes a carriage capacity allocation strategy and optimization method. This method allocates the train capacity to each station by reserving and allocating carriages to ensure that all passengers at each station, especially for the congested stations, can receive fair service and the extreme congestion can be alleviated. Considering the proposed strategy, this studytakes the number of reserved carriages of each train at each station as the decision variable. A linear integer programming model for carriage capacity allocation is developedto minimize the total waiting time of all passengers at each station. A numerical experiment is carried out in Batong Line of Beijing metro for the carriage capacity allocation in the downward direction during the morning peak hour from 7:00-10:40 am on weekdays. The experimental results show that the maximum number of passengers gathered at each station in the line has fallen within an acceptable range, among which the maximum number of gathered passengers has been reduced from 3642 to 1345, a decrease of about 63% . The total waiting time of passengers only increased form 418027 minutes to 420099 minutes, an increase of 0.5%. The results show that the train capacity allocation method performed well in alleviating the extreme congestions in oversaturated metro lines and realized the balance of passenger flow gathering at each station. The method can improve overall operational safety at the line leveland ensure the quality of passenger services.
    References | Related Articles | Metrics
    Migration Law of Urban Traffic Congestion Risk Distribution Considering Uncertainty of Boundary Conditions
    HU Li-wei , ZHAO Xue-ting
    2022, 22(3): 120-131.  DOI: 10.16097/j.cnki.1009-6744.2022.03.014
    Abstract ( )   PDF (4153KB) ( )  
    In order to analyze the influence mechanism of the traffic connection boundary (traffic volume, traffic diffusion coefficient) on the traffic congestion migration law in the study area of a specific city, this paper considers the distribution and migration law of urban traffic congestion risk from a new perspective that it applies the numerical simulation model of pollutant migration to urban road traffic congestion risk migration. First, the urban traffic survey data in a certain area of Guiyang City from June to July 2020 and crawled AutoNavi map data are processed. The study area is divided into Area 1 and Area 2. On this basis, considering the traffic connection boundary and establishing a numerical surrogate model based on urban traffic congestion risk distribution and migration, the Monte Carlo method is used to statistically analyze the input and output results of a study area in Guiyang. The actual road network data values are taken as input into the surrogate model. By comparing the output results with the practical results of the road network and the AutoNavi map, the established model is 3.7% more accurate than the AutoNavi map, and the migration law of simulated traffic congestion risk is higher. The results show that the traffic connection boundary has a great influence on the prediction of the numerical model of urban traffic congestion migration. Statistical analysis of the simulation results in the study area can effectively evaluate the impact of different traffic connection boundary conditions on the risk distribution of traffic congestion and accurately predict the risk of varying levels of traffic congestion within the country. The model established in this paper can efficiently and accurately analyze the riskdistribution and migration law of urban traffic congestion, and it can provide a new reference for urban road network subdivision and sub-point control of traffic congestion.
    References | Related Articles | Metrics
    Distracted Driving Recognition Considering Distraction Types
    ZHOU Yang , FU Rui , LIU Zhuo-fan
    2022, 22(3): 132-139.  DOI: 10.16097/j.cnki.1009-6744.2022.03.015
    Abstract ( )   PDF (1936KB) ( )  
    The recognition of driver distraction states is the basis of driver distraction warning. Because visual distraction poses a greater threat to driving safety than cognitive distraction, this paper studies the recognition of both types of distraction. A simulated driving test was designed for normal driving and different types of distraction. The 1-back task was used to create the cognitive distraction and the phone tasks was used to create the visual distraction of drivers. Then the driving performance, eye movement, and head movement were collected and extracted. The sequential backward selection was used for feature selection, and the grid search method was applied to determine the best time window and model parameters for driver distraction recognition. The results show that the established model based on random forest achieves 94.07% macro accuracy, 93.89% macro recall, and 93.98% macro F1 value on the test set. The classification performance is better than the two traditional methods for comparison, indicating that the model can accurately classify the three states of drivers. From the ranking results of the feature importance output by the random forest model and the classification results of the model with different types of features as input, it is found that driver's eye movement and head movement features are more important for the recognition of different distractions of drivers. This study provides a basis for the driver distraction warning system to determine the risk level according to the type of distraction.
    References | Related Articles | Metrics
    Group-level Random Parameter Spatial Modeling for Road Factors of Taxi Speeding Behavior
    LIU Hai-yue , JIANG Chao-zhe , FU Chuan-yun , ZHOU Yue
    2022, 22(3): 140-146.  DOI: 10.16097/j.cnki.1009-6744.2022.03.016
    Abstract ( )   PDF (1521KB) ( )  
    To thoroughly disclose the relationship between taxi speeding behaviors and road characteristics at segmentlevel, this study extracted large-scale taxi speeding behaviors from GPS trajectories collected in downtown Chengdu. The characteristics of speeding behaviors consist of speeding linear density (SLD) and speeding severity (SR) are sorted in different percentiles. Ten types of road characteristics were selected as underlying contributing factors. To restrict the interference of spatial effects, this study examined the spatial correlation among the characteristics of speeding behaviors and developed three types of spatial models. The models include spatial intrinsic conditional autoregressive model (ICAR), spatial error model (SEM), and spatial lag model (SLM), which are based on the Lognormal prior to the response variables (e.g., SLD and SR). In addition to the spatial correlation, the models were extended to be incorporated with random parameters to capture the unobserved heterogeneity among roadways. The results indicate that the spatial models outperform the traditional model with better goodness-of-fit since all the speeding characteristics are observed with severe spatial correlation. We also found the performance of a certain model varies across the type of response variables. In detail, random parameter ICAR model outperforms others on modeling SLD, while SRs on various percentiles are best fitted by different spatial models. The factors of speed limit, roadwaycross-section, and non-motorized vehicle lane have heterogenous effects on the speeding characteristics. The estimates also indicate that the speed limit, roads without divider, non-motorized vehicle lane, and work zone are significantly associated with SLD. Speed limit, non-motorized vehicle lane, overpass or road tunnel, work zone, and road length are significantly related to SR.
    References | Related Articles | Metrics
    Network-wide Traffic Flow Prediction Research Based on DTW Algorithm Spatial-temporal Graph Convolution
    LIU Yi-cheng, LI Zhi-peng , LV Chun-pu , ZHANG Tao , LIU Yan
    2022, 22(3): 147-157.  DOI: 10.16097/j.cnki.1009-6744.2022.03.017
    Abstract ( )   PDF (2383KB) ( )  
    : To deeply explore the complex temporal and spatial characteristics of traffic flow data and establish their dependence relationship, a new traffic flow prediction model, TSARGCN, based on attention mechanism and residual network, is proposed to improve the prediction accuracy of traffic flow parameters. The TSARGCN slices the input data to realize the time periodicity of data mining by multi-branch modeling. A residual network is introduced to ensure the integrity of information transmission in the network. The DTW algorithm was used to calculate the similarity degree of traffic flow sequence between nodes in the road network in the time dimension, and the concept of a time graph was put forward. Based on the spatial-temporal diagram, graph convolution and time convolution were carried out in each branch combined with an attention mechanism, respectively, to capture the spatial-temporal characteristics of the traffic flow and its dependence relationship, and then the spatial-temporal relationship of the traffic flow data was modeled. Experiments on the open data set PEMSD4 show that, MAE and RMSE of the TSARGCN are 19.24 and 27.09, respectively, which are better than those of ARIMA, CONV-LSTM, and ASTGCN.
    References | Related Articles | Metrics
    Real-time Traffic Flow Parameters Estimation Model Based on Generative Adversarial Network
    YAO Rong-han , WANG Rong-yun , ZHANG Wen-song , YE Jin-song , SUN Feng
    2022, 22(3): 158-167.  DOI: 10.16097/j.cnki.1009-6744.2022.03.018
    Abstract ( )   PDF (2311KB) ( )  
    To effectively allocate the spatio-temporal resources of a road network, it is necessary to estimate the traffic flow parameters in real time. The accurate estimation of traffic flow parameters requires the detailed consideration of the spatio-temporal characteristics of traffic flow in the road network. Based on the generative adversarial network, a real-time estimation model that can capture the spatio-temporal characteristics of traffic flow was formulated, that is, the TSTGAN model. This model included a generator and a discriminator. In the generator, the gated convolutional neural network was used to capture the dynamic spatial characteristics of traffic flow, and the long short-term memory neural network based on the attention mechanism was used to analyze the dynamic temporal characteristics of traffic flow. The discriminator consisted of the gated convolutional neural network and the long short-term memory neural network. The generator and discriminator in the generative adversarial network were trained by an adversarial mode, and the real-time estimated values of traffic flow parameters were obtained. The reliability of the TSTGAN model was validated using the traffic flow data obtained from 12 bayonet devices in Zibo City, Shandong Province, China, and 23 loop detectors in Los Angeles, California, America. The results show that: the introduced spatio-temporal block in the TSTGAN model can effectively extract the spatio-temporal characteristics of traffic flow, and the obtained root meansquare and mean absolute errors decrease by 2.12%~42.41% and 1.66%~40.49%, respectively, compared with those obtained from the existing models, which indicates that the formulated TSTGAN model can improve the estimation precision of traffic flow parameters.
    References | Related Articles | Metrics
    Mining and Calculating Travel Time Based on Classification of Grid Traffic State
    XIE Dong-fan , JIA Hui-di, LI Chun-yan, ZHAO Xiao-mei
    2022, 22(3): 168-178.  DOI: 10.16097/j.cnki.1009-6744.2022.03.019
    Abstract ( )   PDF (2641KB) ( )  
    This paper proposes a travel time calculating method based on the traffic state of the grid by mining trajectory data of taxis. With the gridding of a selected region, this paper utilizes the GPS data of taxis to construct the macroscopic fundamental diagram and to fit the flow- density function of the macroscopic fundamental diagram. By applying the Gaussian mixture clustering method, the traffic states are divided into three categories: free flow, mild congestion, and severe congestion. Through analyzing the travel time of grids with the three traffic states, we found that the travel time of grids reveals different characteristics of the distribution. Specifically, Gamma distribution, Weibull distribution, and lognormal distribution correspond to the states of free flow, mild congestion, and severe congestion, respectively. Furthermore, this paper deduces the probability density model for travel time by approximating the joint probability density distribution of travel time in different grids. The empirical results illustrate that the proposed method can quickly calculate the travel time with specific reliability, and the mean absolute error is between 1% and 16%. The proposed method in this paper can provide technical support for traffic guidance and enhanced navigation.
    References | Related Articles | Metrics
    Evaluation Model of Traffic Organization for Left-turn Prohibition Based on Macroscopic Fundamental Diagram
    LI Bing , YANG Hong-yu , ZHENG Zuo-xiong , FENG Yue , WANG Zheng-hui
    2022, 22(3): 179-189.  DOI: 10.16097/j.cnki.1009-6744.2022.03.020
    Abstract ( )   PDF (3644KB) ( )  
    With the increase of traffic pressure at urban intersections, the prohibition of left-turn vehicles has become an important means to alleviate supersaturated traffic congestion and improve traffic efficiency. The purpose of this paper is to overcome the limitations of the existing left-turn prohibition methods, which mostly investigate only a single intersection, but lack of scientific evaluation basis for left-turn prohibition measures in arterial roads and regions. Based on the advantages of regional traffic state description of Macroscopic Fundamental Diagram (MFD), the evaluation method of the left-turn prohibition traffic organization has been established for different evaluation scopes, such as an isolated intersection, multi-intersections, and regional intersections, respectively. Firstly, the left-turn prohibition flow transfer of intersection is described based on the logit model. Secondly, the MFD after the left-turn prohibition is constructed with the transfer proportion of left-turn flow and the road grade as the key factors, and the evaluation model of the left-turn prohibition traffic organization based on MFD is established. Finally, the reliability of the model is evaluated by an example analysis of the Beijing Road area in Kunming. The results show that the use of left-turn prohibition at an isolated intersection or arterial road can improve the traffic operation efficiency of the road network when the proportion of left-turn traffic is less than 15% in the simulation area; only left-turn prohibition at the isolated intersection can improve the operation efficiency of the road network when the proportion of left-turn traffic is 15%~ 20%; and in any state of left-turn traffic ratio, prohibition left-turn on the whole region will reduce the traffic operation efficiency.
    References | Related Articles | Metrics
    Rental and Allocation of Shared Mixed Parking Spaces Considering Matching Priority
    LI Chang-min , WEI Wen-bin, ZHANG Lu
    2022, 22(3): 190-197.  DOI: 10.16097/j.cnki.1009-6744.2022.03.021
    Abstract ( )   PDF (2171KB) ( )  
    This paper establishes the matching priority based on travelers' demands and in full consideration of the parking and charging resources of mixed parking spaces. A shared parking space rental and allocation model with matching priority (MPRA) is developed to maximize the total income of the shared platform, deducting the penalty cost for rejecting parking requests from rental incomes, with the constraint that the charging amount does not exceed the parking lot load. Based on the matching priority principle and model characteristics, the study proposes the improved ant colony algorithm by introducing a block conflict matrix and the pheromone update strategy. The numerical examples imply the matching priority and the effect of problem scales on MPRA allocation scheme. The results show that the proposed MPRA scheme results in higher charging demand matching rate than the profit maximization scheme, the average improvement rate is 28.96%. This advantage is inversely proportional to the number of travelers or spaces, and the profit improvement rate increases with the increase of the problem size. When the number of parking spaces is scarce, the MPRA scheme can improve the parking utilization rate by 17.87% and improve the charging demand matching rate by 113.96% compared to the first-come-first-serve scheme.
    References | Related Articles | Metrics
    Bike-sharing Trips Spatial Heterogeneity and Driving Factors
    SUN Chao, LU Jian
    2022, 22(3): 198-206.  DOI: 10.16097/j.cnki.1009-6744.2022.03.022
    Abstract ( )   PDF (2025KB) ( )  
    This paper examines the spatial heterogeneity characteristics of bike-sharing trips and the driven factor assessment based on trajectory data mining. The kernel density analysis and hotspot detection are performed to obtain the sampled analysis area and characterize the generation of bike-sharing travel with hot values to reduce the interference of scale effects. The semi-variance function of spatial statistics is introduced to simulate the structural and stochastic variation patterns of bike-sharing travel generation, to explore the spatial heterogeneity characteristics and to determine the scale range of neighborhood effects. The slope of the spatial series is used to characterize the change trend. Meanwhile, the respective driving forces of land use, neighborhood effects and other built environment on spatial heterogeneity are distinguished through the improved spatial lag and residual models. The bike- sharing in Beijing is analyzed as a case study. The results indicate that the spatial autocorrelation of bike-sharing trips in Beijing is moderate and positive, and the spatial heterogeneity is consistent with the exponential model. The decay radius of spatial autocorrelation is 1860 meters, and the neighborhood effect becomes insignificant when the distance is greater than this threshold. The built environment has the most significant impact on the spatial heterogeneity of bike-sharing trips, the neighborhood effect is at an intermediate level, and the land use has the smallest impact on bike-sharing trips.
    References | Related Articles | Metrics
    Urban Logistics Unmanned Aerial Vehicle Vertiports Layout Planning
    ZHANG Hong-hai , FENG Di-kun, ZHANG Xiao-wei, LIU Hao, ZHONG Gang, ZHANG Lian-dong
    2022, 22(3): 207-214.  DOI: 10.16097/j.cnki.1009-6744.2022.03.023
    Abstract ( )   PDF (1919KB) ( )  
    This paper focuses on the layout planning of urban logistics Unmanned Aerial Vehicle (UAV) vertiports. In consideration of different types of logistics UAV vertiports, this paper proposes a vertiports layout planning model with the objective of minimizing the total economic cost and maximizing the customer satisfaction. The constraints of the model involve no-fly zone, UAV performance, vertiport capacity, and other factors. The human learning optimization algorithm (HLO) is designed and the random learning operator, individual learning operator and social learning operator are introduced in the algorithm to solve the model. The simulation experiment is then performed with real geographic data and logistics data to verify the effectiveness of the model and algorithm. The experimental results show that the proposed model can generate reasonable layout planning of vertiports, which is suitable and effective for large-scale resource allocation problem. The HLO algorithm shows better solution accuracy and convergence speed than the genetic algorithm (GA) The parameter analysis shows that the optimal economic cost weight is 0.4 and the optimal customer satisfaction weight is 0.6 based on the simulation environment. The optimal algorithm learning probability parameters are 5/n and 0.8+2/n. The study results could provide decision-making support for the layout planning of the actual urban logistics UAV vertiports.
    References | Related Articles | Metrics
    Modeling Towards Freeway Real-time Traffic Crash Prediction Considering Multi-dimensional Dynamic Feature Interactions
    YUAN Zhen-zhou , HU Yan-ran , YANG Yang
    2022, 22(3): 215-223.  DOI: 10.16097/j.cnki.1009-6744.2022.03.024
    Abstract ( )   PDF (1700KB) ( )  
    This paper investigates the impact of weather, road features, and the dynamic mutual interactions among traffic flow, weather, road, and time on the accuracy of real-time crash risk prediction. The study developed four datasets based on the crash data, traffic sensor data, weather data, and road data collected from the Beijing section of the Beijing-Harbin Freeway. The datasets include (1) the simple traffic flow data; (2) the combined traffic flow, weather, and time data; (3) the combined traffic flow, road, and time data; (4) combined traffic flow, weather, road, and time data. By considering the interactions of multi-dimensional dynamic features, this study proposes a real-time crash risk prediction model based on the Deep & Cross Network (DCN). The results demonstrate that the DCN model achieves higher accuracy than other methods in real-time crash risk prediction. The Area Under Curve (AUC) of the model is 0.8562 and the proposed model is able to correctly classify 84.26% of non-crash data and 77.55% of crash data with the probability threshold of 0.2. The DCN model used in this study can effectively predict the occurrence of freeway crashes and collisions in time, under the condition of multi-dimensional dynamic feature interactions. The proposed method has great potential to support the freeway safety management departments of China in both theoretical and technical aspects.
    References | Related Articles | Metrics
    Heterogeneous Effects of Built Environment on Elderly Traffic Accident Injury Severities
    PAN Yi-yong, WU Jing-ting, SHI Ying, MIAO Xuan-ye
    2022, 22(3): 224-230.  DOI: 10.16097/j.cnki.1009-6744.2022.03.025
    Abstract ( )   PDF (1439KB) ( )  
    This paper uses a random parameter approach with heterogeneity in means and variances to analyze the heterogeneity of elderly traffic accident injury severities. The 2019 traffic accident data in a state of the United States is used for the analysis. 17 influencing factors are selected based on the characteristics of elderly, vehicle, road, road environment, and built environment. The average elasticity coefficient is used to reflect the effect of each factor on the accident injury severities. The results show that the goodness of fit of Logit model based on mean and variance heterogeneity random parameter is better than traditional method. Among the built environment factors, the existence of a shopping center in the 300-meter buffer zone was a random variable. Its mean value was significantly correlated with the male elderly and the number of lanes with the coefficient of 3. The variance was significantly correlated with the uncontrolled road control mode. In addition, motorcycle, tangent segment, wet road surface and other factors significantly increased the severity of injury severities. The results of this study provide refences for policy makers to adopt appropriate strategies to reduce the severity of elderly traffic accident injuries.
    References | Related Articles | Metrics
    Inferential Analysis of Vehicle Accident Severity in Mountainous Highway Crossing Villages
    JI Xiao-feng, ZHAN Huan-qin, PU Yong-ming, QIN Wen-wen
    2022, 22(3): 231-237.  DOI: 10.16097/j.cnki.1009-6744.2022.03.026
    Abstract ( )   PDF (1534KB) ( )  
    This paper investigates the impact of roadway alignment, driver attributes, vehicle types, accident patterns and other main factors on the severity of vehicle accidents on mountainous highway that cross villages and towns. Based on the 2012 to 2017 accident data of Yuanshuang Highway (Yuanmou-Mouding) in China, this paper identified 15 major factors from the aspects of human, vehicles, roads and environment using the social network analysis method. The Bayesian network model is developed based on machine learning method. Taking accident severity as decision variable, the study analyzes the reasoning results of the interaction between different evidence variables and driving behaviors. The results indicate that the combination of unsafe driving behavior and some risky factors would increase the severity of accidents. When trucks were involved, the accident rate with injury would increase by 8.2% if the safety distance is not satisfied. In the curve segment with slopes, the accident rate with injury would increase by 19.6% if the driver misjudges the driving conditions. The accident rate with injury would also increase by 5.4% on rainy days and the driver misjudges the driving conditions. The probability of injury would increase by 3.1% in the rollover accident caused by improper operation of the driver.
    References | Related Articles | Metrics
    Robust Optimization Model and Algorithm for Maritime Inventory Routing Problem of Multi Fuel Products
    YANG Hua-long , WU Yan-hua, SUN Yi-lun
    2022, 22(3): 238-246.  DOI: 10.16097/j.cnki.1009-6744.2022.03.027
    Abstract ( )   PDF (1869KB) ( )  
    This paper studied the maritime inventory routing problem (MIRP) of multi fuel products with uncertain demand and unknown probability distribution. A distribution strategy was proposed for undedicated compartments and non-fixed calling supply port. A nonlinear robust optimization model of MIRP was established by setting up the uncertain demand budget threshold of multi fuel products for the cumulative voyages with given supplier's conservative coefficients, where the total cost of supplier's fuel inventory and distribution was minimized. An improved hybrid adaptive genetic algorithm (HAGA) was designed to solve the model. The numerical examples indicate that the distribution with undedicated compartments and non- fixed calling supply port can effectively reduce the total cost of supplier's fuel inventory and distribution. There exists a different right supplier's conservative coefficient value for each type of fuel product at each demand port, for example, the change in customer service level tends to flatten when the supplier's conservative coefficient exceeds the right value. This study provides a useful reference for suppliers' decisionmaking on multi fuel product MIRP.
    References | Related Articles | Metrics
    Extended Belief Rule Based System for Risk Prediction of Pirate Attacks
    LV Jing , QI Hai-di, LI Bao-de
    2022, 22(3): 247-254.  DOI: 10.16097/j.cnki.1009-6744.2022.03.028
    Abstract ( )   PDF (1777KB) ( )  
    To effectively predict pirate attacks and reduce possible losses, this paper proposes a pirate attacks risk prediction model based on the joint optimization of extended belief rule base (EBRB). By introducing the Relief F algorithm and differential evolution algorithm, this study optimizes the EBRB system from the perspective of structure and parameters, which ensures the EBRB system has the optimal number and value of parameters. The real pirate event data set is used for the case study. The results show that the reasoning results of the jointly optimized EBRB system match well with the actual situation. Compared with the initial EBRB system, the proposed method improves the risk prediction accuracy of pirate events by 60% . In addition, compared with the traditional prediction models, the prediction model based on the joint optimization of EBRB shows certain advantages in improving the prediction accuracy.
    References | Related Articles | Metrics
    Modeling and Simulation of Interaction Between Road Users at Mixed-traffic Intersections
    LI Yi-xin, NI Ying , SUN Jian
    2022, 22(3): 255-266.  DOI: 10.16097/j.cnki.1009-6744.2022.03.029
    Abstract ( )   PDF (2717KB) ( )  
    To accurately describe the interaction and microscopic motion of road users, and provide traffic flow for the virtual testing of autonomous vehicles, this paper develops an interaction model for representing interactions between road users at mixed-traffic intersections. Based on the human cognitive process, the model is designed as a threelayered“perception-decision-action”framework to capture the whole interaction process of road users in mixed traffic. In the perception layer, an interaction selection model is proposed to dynamically decide the interactive road user based on their trajectories. The Decision layer chooses a suitable behavior decision result for road users according to the traffic context. The action layer calculates the behavior trajectory, acceleration, and steering angle based on the decision result to jointly capture the two- dimensional motion of road users. Then, a virtual testing platform is established to verify the performance of the proposed model using the Opendrive road network. Simulation results show that the proposed model can well represent the interaction between different kinds of road users, including motor vehicles, bicycles, and pedestrians. This study provides a way for exploring how to provide traffic flow and satisfy the high precision simulation requirement for virtual testing tools.
    References | Related Articles | Metrics
    Urban Expressway Merging Area Simulation Model Based on Heterogeneity of Driving Behavior
    SHUAI Bin , MI Rong-wei, ZHANG Rui, LEI Yu
    2022, 22(3): 267-275.  DOI: 10.16097/j.cnki.1009-6744.2022.03.030
    Abstract ( )   PDF (2443KB) ( )  
    To investigate the law of vehicle operation in the merging area of urban expressways, this paper proposes a cellular automata simulation model of the confluence area considering the heterogeneity of driving behavior based on vehicle natural trajectory data. The model divides the confluence into the upstream area, merging area, downstream area, and these three areas are composed of 11 roadway sections. This paper first uses the Kalman filter algorithm to denoise the natural trajectory data, and then calculates the driving behavior characteristic parameters of each vehicle. The K-means clustering analysis is also performed and the clustering effect evaluation index Silhouette coefficient is introduced to divide driving behavior into four typical types: conservative-cautious, aggressive-cautious, conservativereckless, and aggressive- reckless. Based on the classification results, the paper develops a car- following model that considers acceleration and random slowdown probability heterogeneity and a multi-level heterogeneous lane-changing model that considers the safety distance and decision-making of lane changing behavior. In the context of each space occupancy rate, the model is simulated using Matlab software, and the traffic volume, density, speed, space-time position, lane changing frequency and other parameters of the lanes in the merging area are analyzed under the conditions of homogeneous driving behavior and heterogeneous driving behavior. The simulation results show that when the space occupancy is 10% to 20% , the homogeneous traffic flow is more likely to produce local traffic congestion and traffic flow failure conditions than the heterogeneous traffic flow. The peak traffic volume of thehomogeneous traffic is 27.1% lower than the heterogeneous traffic volume. With the increase in space occupancy, the driving frequency of homogeneous vehicles and heterogeneous vehicles both show an increase-steady-decrease trend, while the maximum value of lane-changing frequency for heterogeneous driving behavior is 20.74% higher than that of homogeneous traffic flow
    References | Related Articles | Metrics
    Cellular Automata Model of Multi-lane Weaving Area Based on Lane-changing Probability Distribution
    XIE Ji-ming , PENG Bo , QIN Ya-qin
    2022, 22(3): 276-285.  DOI: 10.16097/j.cnki.1009-6744.2022.03.031
    Abstract ( )   PDF (2894KB) ( )  
    The interweaving area is an important part of the expressway system, which is prone to cause traffic bottlenecks due to frequent vehicle lane changes and complex interactions. In this paper, we extract high precision vehicle trajectories with a time resolution of 0.1 s and a spatial resolution of 0.1 m· px-1 in an urban multi-lane interweaving area, then analyze the traffic flow and vehicle behavior characteristics in the interweaving area and adjacent road sections and propose a partitioned cellular automata model. In the upstream and downstream lane change models, the lane change motivation rule, spacing rule, and logistic lane change probability rule based on speed difference and vehicle spacing are established. For the interweaving influence zone, the lane change motivation rules are established considering speed, spacing, and path transition requirements, multi-step decision rules for lane change timing are constructed based on safety risks, and lane change probability rules are proposed based on the Gaussian distribution model of lane change frequency. The sensitivity simulation tests are conducted to analyze the key parameters, which show that the model has the potential for practical application in evaluating different lane change states in the interweaving area. Simulation and actual measurement show that the characteristics of flow, speed, and density, and lane change distribution of the model are consistent with the reality, which can effectively reflect the lane change demand and intensity variability of vehicles at different locations and portray the complex following and lane change behaviors in multi-lane interweaving areas.
    References | Related Articles | Metrics
    Car-following Model with Adaptive Expected Driver's Following Distance and Behavior
    NI Jie , ZHANG Kai-duo, LIU Zhi-qiang, GE Hui-min
    2022, 22(3): 286-292.  DOI: 10.16097/j.cnki.1009-6744.2022.03.032
    Abstract ( )   PDF (2005KB) ( )  
    To satisfy the personalized demand of intelligent vehicles and to improve the satisfaction and acceptance of intelligent vehicle human-computer interaction, a two-layer driver car-following model framework was constructed, and a personalized driver car-following model was proposed. The models can adapt the driver's expected following distance and behavior. Firstly, the equilibrious car-following data was extracted. The first layer model was established by using the Gaussian Mixture Model (GMM) and Probability Density Function (PDF), which was the driver's expected following distance model. Then, the expected following distance parameter was introduced into the model, and the driving behavior was learned based on Gaussian Mixture Model (GMM) and Hidden Markov Model (HMM). The second layer model, i.e., the personalized car-following model, was established to predict the acceleration of future time. Next, the effect of different numbers of GMM components on the model performance was studied, and the comparison was made among the two-layer driver car-following model, the Gipps model, the optimal distance model (ODM), monolayer model and the general model. Finally, the results of the 8 drivers' naturalistic driving behavior data show that the number of GMM components is positively correlated with the model performance. Under the optimal Gaussian model component, the mean predictive deviation of 8 drivers in the training set is 0.101 m·s-2 , and 0.123 m·s-2 in the test set. The model calculation results of randomly selecting one of the drivers' experimental data show that the mean absolute deviation of acceleration is 0.087 m·s-2 and 0.096 m·s-2 , and the prediction results are better than that of the Gipps model, the ODM model, monolayer model and the general model by more than 30% , which is moreconsistent with the actual car-following behavior of the driver
    References | Related Articles | Metrics
    Simulation of Highway Traffic Bottleneck Via Cellular Automata
    LV Wei, HUANG Guang-chen , WANG Jing-hui
    2022, 22(3): 293-302.  DOI: 10.16097/j.cnki.1009-6744.2022.03.033
    Abstract ( )   PDF (3469KB) ( )  
    To analyze the congestion caused by road bottlenecks on the highway, this paper uses the improved KernerKlenov-Wolf (KKW) model to establish car-following rules, and comprehensively considers the impacts of vehicle spacing and speed on vehicle lane-changing behaviors to establish free lane-changing and mandatory lane-changing rules. This paper also analyzes the distribution of the congestion area upstream of the road bottleneck, the characteristics of lane-changing behaviors, and the changes of traffic parameters on the lane through simulations under different traffic flow conditions. The results show that: with given traffic volume, the length of congested area in merging lane is in a dynamic equilibrium state and will not change with time. In addition, the convergence behavior upstream of the road bottleneck can lead to a severe speed drop on the target lane. The vehicle speed changes in the convergence lane and the target lane tend to be consistent. From the perspective of lane-changing cluster characteristics, the low-speed merging vehicles tend to change lanes in small groups in the upstream of the road bottleneck due to heavy traffic volumes, which would cause severe traffic shocks in the target lane. After the bottleneck disappears, the traffic recovery time increases linearly with the increase of the entering traffic flow.
    References | Related Articles | Metrics