The carbon emission of the traffic and transportation industry accounts for 10% of the total emission throughout the country. The industry, which provides support for the development of national economy, has its own responsibility of carbon emission reduction in the meantime, and is one of the important arenas of carrying out carbon peaking and carbon neutrality strategy. With the subject of Tasks and Measures of Carbon Emission Reduction for China's Traffic and Transportation Industry in the New Period, and taking the strategy of carbon peaking and carbon neutrality as the flag, it is specified in the session both tasks and key fields of carbon emission reduction, based on making fully use of comparative advantages of each transportation mode. It is also discussed in the session the measures and feasible roadmaps in key fields, with comprehensive considerations of social and economic development in the new period, substitution degree of new energy products, and economic feasibility of new technologies.

This paper addresses the problem of optimizing cold chain multimodal transportation routes under uncertainty of carbon trading prices. Triangular fuzzy numbers are employed to describe the carbon trading price, taking into account the characteristics of price fluctuations in the current market. The risk preference level of multimodal transport operators and the customer's soft and hard transport time window requirements are combined to establish the fuzzy credibility function for the carbon trading price and the customer satisfaction function, respectively. Accordingly, a bi-objective optimization model considering transportation cost and customer satisfaction is constructed, and a fuzzy adaptive non-dominated sorting genetic algorithm (FANSGA-II) is designed. The numerical example results show that compared with the random uniform distribution method, the triangular fuzzy number method can enhance the stability of the optimization results and reduce the coefficient of variation of transportation cost from 20.3% to approximately 4.4%. The results of the sensitivity analysis indicate that the level of risk preference of multimodal transport operators is positively correlated with the proportion of railroad and waterway transport modes. Furthermore, the implementation of a carbon trading policy is predicted to result in a reduction in carbon emissions by at least 7%. The lower the refrigeration temperature of the cargo, the more sensitive the carbon trading risk preference value is to the choice of transportation mode. However, once the risk preference value reaches a specific value, the chosen transportation mode will no longer change. The findings of the study can provide a reference for multimodal transport operators to develop cold chain transportation options.

The New Western Land-Sea Corridor encompasses multiple ports and stations, rendering it vulnerable to risky events that can trigger cascading failures through goods transfer across nodes. This paper first constructs sub networks based on distinct transportation modes and integrates geographically adjacent nodes to form a comprehensive transportation network. Subsequently, it identifies critical nodes contributing to network vulnerability using an enhanced CRITIC-TOSIS method. A nonlinear load-capacity cascading failure model is then developed to assess network vulnerability changes through connectivity and overload efficiency indicators. A vulnerability simulation of the network is conducted to compare and analyze the impact of various parameters and strategy settings on network vulnerability. The results reveal that: (1) the composite transportation network significantly reduces vulnerability compared to individual sub-networks under attack; (2) an intentional attack strategy focusing on node vulnerability leads to rapid network collapse, with a 99% decrease in the vulnerability index upon the failure of 10 nodes; (3) adjusting capacity coefficients within a specific range can effectively mitigate network vulnerability. When the capacity redundancy coefficient α≥4 , β≥1.4 and overload capacity coefficient γ≥0.3, the network exhibits robust resilience against diverse attack strategies, maintaining a low vulnerability level.

Developing new energy vehicles in the current era is a crucial step toward enhancing energy conservation and reducing emissions in China's transportation sector. This paper employs a Cournot evolutionary model that takes market share into account to analyze the competitive dynamics among new energy vehicle manufacturers under the dual-credit policy, offering recommendations based on the findings. The study identifies several key points: first, the necessary condition for halting dual-credit trading is the internal credit balance among internal combustion engine (ICE) vehicle manufacturers. The dual-credit policy will officially conclude once the goal of eliminating negative credits for ICE manufacturers is realized. Second, during the initial phase of credit trading, the policy tends to favor less competitive new energy vehicle manufacturers, resulting in a "bad money drives out good" phenomenon, where low-quality products dominate the market and lead to regulatory failures. As ICE manufacturers enter the market, the marginal effect of the dual-credit policy on promoting the growth of the new energy vehicle market diminishes, and the competitive landscape stabilizes based on the varying levels of competitiveness among all manufacturers. Third, the cyclical evolution of the dual-credit market influences new energy vehicle manufacturers, transitioning from a positive impact that offsets production costs in the early phase to a negative impact that constrains capacity growth in the later phase. For ICE manufacturers, the policy shifts from initially undermining their competitive advantage to later promoting their transformation and development. Ultimately, the disappearance of bonus from the credit policy will compel new energy vehicle manufacturers to prioritize technological innovation, thereby enhancing their core competitiveness and management capabilities, and fostering high-quality development within the transportation sector.

The lane-changing behavior of vehicles approaching intersections will constrain the improvement of intersection traffic efficiency. Based on this, this paper proposes a Segmented Cooperative cOntrol Method for Urban Road Traffic Flow (SCOM-URTF) in a connected vehicle environment, which adopts a bi-level optimization model to achieve dynamic division of road section functional zones and collaborative optimization of traffic flow between road section and intersection. The upper-level model designs a Misaligned Lane-changing with Separated Lane Speed Guidance (ML-SLSG) to promote rapid lane changes for left and right turning vehicles through rearranging the vehicles entering from upstream intersection in longitudinal space, minimizing the vehicle lane-changing zone length, and balancing lane group traffic flow. The lower-level model uses dynamic programming to optimize the trajectory of connected vehicles and intersection signal timing parameters with the goal of minimizing average vehicle delay. The simulation results show that ML-SLSG can effectively shorten the total length of lane-changing. At the same time, the longitudinal trajectory optimization model proposed in this paper can reduce average vehicle delays at intersections by 5.9% ~8.0% under low, medium and high traffic demands. And after further collaborative optimization of vehicle trajectory and signal timing, the average vehicle delay can be further reduced by 3.7%~22.8%. Comparative studies with similar methods have shown that SCOM-URTF is more suitable for traffic environments where multiple driving behaviors are coordinated with each other. Sensitivity analysis shows that higher connected and automated vehicle penetration rates and road speed limits can help reduce average vehicle delays, and increasing the spacing between intersections can initially reduce average vehicle delay, but there may be a delay rebound after reaching the critical point. However, the coordinated optimization of trajectories and signals can effectively curb the rebound of delays.

A novel dynamic control logic and method for optimizing the Right-of-Way of connected vehicles utilizing urban bus lanes is proposed to address issues such as low lane utilization and poor performance under high traffic density. This approach ensures priority for buses and dynamically identifies and enhances the spatiotemporal resource utilization in bus lanes. A method for identifying the remaining spatiotemporal resources in bus lanes is defined, and a vehicle travel time calculation method is designed, taking into account the vehicle kinematics and driving behavior, including right-turning vehicles. Based on predefined rules for accessing bus lanes, a candidate set of vehicles eligible for lane borrowing is identified, and a road rights allocation optimization model is constructed, allowing selected connected human-driven/automated vehicles to utilize bus lanes. This enables effective utilization of the remaining spatiotemporal resources in bus lanes. The proposed method is validated through simulation experiments using a road in Jinan City, with parameters such as traffic flow intensity, bus arrival intervals, and proportion of right-turning vehicles matching the actual road conditions. The results demonstrate significant improvements in optimization objectives compared to traditional bus lane usage methods on the actual road and the "bus lane with intermittent dynamic priority" method proposed in the existing literature. Compared to the actual and literature-based control methods, the proposed method reduces delays for non-right-turning vehicles by 30% and 16%, and for right-turning vehicles by 24% and 26%, while ensuring bus priority. Moreover, the optimization effectiveness of the proposed method becomes more significant with increasing traffic intensity.

Urban rail transit systems often experience operational disruptions or reduced service capacity during peak hours, major events, and adverse weather conditions. To effectively mitigate the negative impacts of these disruptions on passenger flow and enhance the resilience of urban rail transit systems, this paper proposes an optimization method for passenger flow guidance in response to disruptive events. First, considering the impact of disruptive events and the compliance rate of passenger guidance, this paper develops a rail transit passenger flow guidance model with the goal of minimizing the total travel time of passengers in the system. Then, a column generation-based exact algorithm is designed, and Gurobi is used to solve the restricted master problem. The A* algorithm is applied to solve the pricing subproblem, and the branch-and-bound algorithm is utilized to find integer solutions. Through actual case analysis, it is found that the acceleration strategies designed in this paper can improve the solving efficiency by 66%~89%, with performance significantly superior to using Gurobi alone. Simulations of scenarios ranging from minor to severe disruptions demonstrate that the proposed optimization method is applicable to urban rail transit passenger flows of varying scales, effectively guiding passenger travel paths under various disruption intensities.

To solve the cross-phase conflict problem in traffic flow at signalized roundabouts, this paper analyzes the vehicle trajectory within roundabout and develops a fundamental cycle equation and cross-phase interval equations based on the characteristics of phase passage time, phase interval time, and minimum phase interval time. Based on the inequality constraints of phase passage time, phase interval time and signal cycle, this paper proposes a basic signal timing model for roundabouts with the minimzed signal cycle and considers the possibility of negative total loss time in the phase design of roundabouts. The general cycle equations and inequality constraints are analyzed under different ordered phase sets. The variation of the cycle range with the flow rate of each phase are analyzed under different phase sets, and a calculation method is proposed for signal cycle range at roundabout. The simulation verification was conducted using a total flow ratio of a roundabout equal to 1.05 as an example. The results show that when signal cycle was set to the shortest signal cycle of 52 seconds, all evaluation indicators reached or approached the optimal value. The proposed method was compared with the two-stop-line-for-left-turn method and the reconstruction into a cross-intersection under three different traffic volume of entrance. It was found that the delay time and the average queue length of two roundabout control methods was reduced by 52% and 55%, compared to the cross-intersection control method. The average delay and the average queue length of the proposed method was reduced by 43% and 44%, compared with the two-stop-line-for-left-turn method, when the flow at the opposite entrance was unbalanced. In addition, taking the Huaxia-Jinsui roundabout in Guangzhou as an example, simulation analysis showed that the proposed method had the overall control benefits in terms of vehicle delay, stops and queue length.

Passenger overcrowding during peak hours is a critical challenge in urban rail transit of large cities. It not only increases the risk of passenger stampede accidents within stations but also leads to significant differences in waiting times between upstream and downstream stations. Considering the stochastic and dynamic nature of passenger demand, this paper develops a multi-objective passenger flow control optimization model that aims to enhance travel efficiency and improve passenger travel fairness. The model can determine reasonable boarding ratio of passengers at each station, and optimize the average waiting time of all passengers (efficiency) and the Euclidean distance between the average waiting time of passengers at each station (fairness) and the ideal value under the feasibility constraints. To solve this multi-objective stochastic optimization model, an online optimization algorithm is developed to make passenger flow control decisions for each demand scenario. Numerical experiments based on passenger flow data from Beijing Metro Line 5 show that, compared to the benchmark first-come-first-served policy, the proposed method can significantly improve operational efficiency and fairness. Moreover, the algorithm achieves results close to those obtained by direct solving with Gurobi while reducing the computation time by 76.3%. This paper provides an effective strategy and methodological reference for addressing the problem of passenger overcrowding in urban rail transit during peak hours.

To effectively deal with the impact of multi-disturbances (i.e., short-time section blockage, section speed limitation, and extended station dwell times) on metro operations, an integer linear programming model for integrated train timetabling and rolling stock rescheduling is built, with five rescheduling strategies (i.e., holding, cancellation, short-turning, skip-stop, and rolling stock deadheading to/from the depot) employed to minimize the weighted-sum of train delays, passenger revenue loss, number of operated backup rolling stocks, and rolling stock deadheading distance. The constraints of train operation, rolling stock circulation, and turn around are considered. Then, the model is solved by an iterated local search algorithm. The proposed model and algorithm are tested on a set of instances that are constructed based on the real-life dataset of a metro line. The results show that: (1) For different cases of multi disturbances, the proposed method can generate integrated rescheduling solutions within a short time, where the average computational time is 15.55 s; (2) The proposed strategies of cancellation and short-turning can control the number of delayed trains and prevent delays from spreading, where the strategy of rolling stock deadheading to/from the depot can reduce secondary delays resulted from restricted capacity at turnaround stations; (3) Compared with fixing rolling stock circulation plan and ignoring the cost of rolling stock rescheduling, the proposed rescheduling method can respectively reduce the total train delays by 74.9% and 7.3%, where the number of operated backup rolling stocks and rolling stock deadheading distance are respectively reduced by 23.1% and 20.1% after considering the cost of rolling stock rescheduling.

The through operation mode is an important part of the integrated regional rail transit system. This paper introduces the minute-dependent passenger flow as the input, takes into account the common benefit of multiple participants, and examines the relationship between the departure time of suburban and the through operation of subway. Considering the constraints of train departure interval and train service capacity, this paper develops an integer programming model from the perspectives of subway operation income and passenger travel quality to optimize subway through operation scheme and suburban timetable. Based on the characteristics of the model, two sets of logic variables are introduced to reconstruct the model. The constraint method is used to transform the model into a single objective optimization model, which is accurately solved by commercial optimization solver Gurobi. Several groups of numerical examples were analyzed based on two rail transit lines. The results show that the through operation of subway can not only save the travel time of passengers but also increase the income of operating enterprises. The proposed subway through operation scheme can increase the revenue of the subway operation company by 18000 RMB. The adjusted suburban timetable can save 8.3% of passenger waiting time, which can better match the spatial and temporal distribution of passenger demand.

Based on the 2018 Shanghai bike-sharing order data, this paper uses the multi-scale geographically weighted regression (MGWR) model to analyze the influence mechanism between the metro built environment and the bike sharing trips. It investigates the heterogeneous characteristics of the influence effects from two major attributes of the metro station type and spatial location. The results show that there are four different types of impacts on the attraction and generation of bike-sharing trips, including, all positive, all negative, double positive or negative, and positive and negative opposite. The main factors influencing bike-sharing trips vary significantly between different metro stations, and this variation becomes even more pronounced when considering trip generation. The distance between metro stations and the city center has a significant impact on the ranking of the key factors while there is marginal difference between transfer stations and non-transfer stations. The scenarios of "Metro Station Density > Population Density > Bus Station Density" and "Residential Land > Metro Station Density" are the most influential combinations of trip attraction and generation within 10 km of the city center. The scenarios of "Population Density > Bus Station Density > Residential Land" and "Land-Use Information Entropy > Population Density" correspond to the most important combinations of trip attraction and generation outside 10 km range from the city center. As the distance of metro stations from the city center increases, the influence coefficients of each factor show different gradual change patterns, such as "<" type, "几" type and "S" type. This indicates that the impact of the built environment on bike-sharing change with the spatial location and there is a complex non-linear change, it is appropriate to adopt a place-specific strategy for bike-sharing around metro stations.

Accurate prediction of bike-sharing trip distributions is critical for urban non-motorized traffic planning and bike-sharing operation scheduling. This paper takes residents' travel destination decision-making behavior as a starting point and proposes a single-factor prediction model for bike-sharing trip distribution considering POI. Based on this model, a two-factor prediction model considering congestion index and its improvement model are developed. Based on bicycle riding data from Futian District, Shenzhen, this study analyzes the clustering phenomenon within the travel OD network. It utilizes a community detection algorithm from complex networks to partition Futian District into four traffic analysis zones (TAZ). The influence of POI, congestion index and travel distance on bike sharing are then analyzed, and it is found that the number of POIs shows a linear positive correlation with the amount of bike sharing trips. At the same time, the congestion index shows a significant positive correlation with the amount of bike sharing trips, especially in the areas with larger travel volumes, whenever the congestion index increases by 0.1, the proportion of bike sharing trips will increase by 6%~7%. The travel distance shows a long-tailed logarithmic distribution characteristics. The prediction results show that during weekends, the accuracy of the two-factor improvement model developed in this paper is 81.2%, 79.5%, 80.1%, and 78.9% for the four TAZs, respectively. During weekdays, the accuracy rates were 78.7%, 76.3%, 80.8%, and 75.5%, respectively. Compared to the radiation model, the prediction accuracy was improved by a maximum of 51.1%.

To address the challenges of low predict accuracy, complex hyperparameter optimization, and limited model interpretability in conventional travel mode choice models and machine learning models, this paper introduces the genetic algorithm and Bayesian optimization for hyperparameter optimization of the extreme gradient boosting machine model (XGBoost). Additionally, the SHAP (SHapley Additive exPlanations) model is integrated to visualize the nonlinear relationship between travel mode attributes and individual characteristics in the choice probability. The proposed model is trained using 5-fold cross-validation to prevent overfitting and is evaluated using Swissmetro dataset to demonstrate its superiority. The results indicate that enhancing the nonlinear representation of the utility function in discrete choice models improves model prediction performance, yet falls short compared to machine learning models. The optimized XGBoost model, employing genetic algorithm and Bayesian optimization, outperforms conventional multinomial Logit models with linear or nonlinear utility functions, as well as standard random forest and non optimized XGBoost models in terms of accuracy, recall, and F1 score for travel choice predictions. The XGBoost model optimized by genetic algorithm exhibits the highest prediction accuracy of 0.781, surpassing models based on conventional multiple grid search. Moreover, hyperparameter optimization using genetic algorithm reduces training time by 81.4% compared to multiple grid search. Furthermore, the study reveals that the cost and time associated with different travel modes significantly influence the choice preferences, with trains and cars being more sensitive to time while the Swiss metro is more sensitive to cost.

This paper investigates the impact factors and nonlinear effects of charging behavior for electric ride-hailing vehicles in the built environment which is an important part of public charging station planning and operation. Based on the ride-hailing order data from Nanjing city, this paper develops an algorithm to identify charging behavior for electric ride-hailing vehicles and proposes ten built environment indicators from five dimensions: density, diversity, design, destination accessibility, and proximity to public transportation. The key impact factors on the charging behavior of electric ride-hailing vehicles are identified using the XGBoost model and Shapley additive explanation (SHAP) value algorithm, and the potential nonlinear relationships between these factors and charging behavior are further analyzed. Additionally, the model's fitting performance is compared with Random Forest (RF) and LightGBM to validate the effectiveness of the XGBoost model in regression fitting. The results show that the XGBoost model has better performance compared with traditional models, with smaller prediction error fluctuation and higher R2 (0.446) than the traditional models. The number of restaurants, distance from the city center, and the number of leisure and entertainment facilities are found to have the most significant impacts on charging behavior. Moreover, all built environment factors show nonlinear effects on the charging behavior, with the distance to the city center showing apositive impact at the beginning and then becomes negative, while other variables exhibit a negative impact at the beginning and then becomes positive.

To meet the escalating demand for sophisticated urban traffic demand management measures, the introduction of a Travel Reservation Strategy (TRS) in managing urban road congestion is expected to play an important role in shaping the future of intelligent transportation systems. However, existing research on TRS frequently encounters challenges, including oversimplified assumptions about total reservation volumes on reserved roads and homogeneous travel choices across users, without adequately considering the multi-dimensional decision variables of individual travel and the complexities of integrating urban multimodal transportation. In this study, the road capacity distribution function was estimated using censored data models and the product limit method, and the sustained flow index was introduced to determine the optimal reservation volume for designated roads. Additionally, considering user heterogeneity, a comprehensive model for multi-user, multi-criteria, and multi-modal traffic mode split and traffic assignment is formulated within the urban multimodal transportation framework. The findings reveal that the optimal reservation volume falls within a range of approximately 79% to 89% of the actual road capacity. Following the implementation of TRS, notable improvements were observed, with average road network speed increasing by 7.6%, average saturation enhancing by 7.9%, and total travel cost decreasing by 1.6%, compared to pre-implementation levels. Notably, the proportion of private car users declined by 4.19%, while the share of public transportation users grew by 3.19% . Heterogeneous travelers with different time values demonstrated distinct responses to TRS, highlighting the need for tailored strategies. These findings provide valuable insights into the potential benefits and challenges of TRS implementation, providing policymakers with essential theoretical underpinnings and contributing to the development of more scientific and effective traffic demand management strategies.

Multimodal transportation leverages the advantages of various transport modes, contributing to cost reduction and efficiency improvements in freight logistics, with route decision-making being a critical factor. The organization of multimodal operations and external environmental changes can cause fluctuations in transportation times. This study considers the impact of stochastic transportation times and transfer times on route optimization in multimodal transportation by introducing trapezoidal fuzzy numbers to represent time uncertainty. A time-window constrained multimodal transportation route optimization model is constructed with the objectives of minimizing transportation costs, carbon emissions, and transportation time. Based on fuzzy chance-constrained programming theory, the uncertainty model is transformed into a more tractable mixed-integer programming model. The evolutionary process is divided into two stages based on the real-time state of the population: the first stage focuses on optimizing the objective function, while the second stage objective optimization with constraint satisfaction. On this basis, a multi stage multi-objective evolutionary algorithm is designed to solve the model. Finally, a case study of a multimodal transportation network demonstrates that the proposed method effectively generates a set of route optimization solutions under uncertain transportation times, with chance constraint satisfaction probabilities exceeding 90% . Compared to the state-of-the-art constrained multi-objective evolutionary algorithms, the hypervolume indicator improves by 2.11% to 41.95%, showing significant performance gains and providing effective route decision-making support for multimodal transportation operators.

To reveal the impact of information from multiple preceding and following vehicles on the stability and safety of complex mixed traffic flow, this paper constructs a car-following model for connected autonomous vehicles (CAVs) and connected human-driven vehicles (CHVs) that considers information from multiple preceding and following vehicles. The model is used to study the stability and safety of complex mixed traffic flow composed of human-driven vehicles (HDVs), autonomous vehicles (AVs), CAVs, and CHVs. Firstly, a complex mixed traffic flow model considering information from multiple preceding and following vehicles is established, and all car-following modes as well as the proportional relationships among the four types of vehicles are analyzed. Secondly, the stability criteria for complex mixed traffic flow under different penetration rates of connected vehicles are theoretically analyzed. Finally, a numerical experiment is designed to analyze the influence of connected vehicle penetration rate and information from multiple preceding and following vehicles on the stability and safety of complex mixed traffic flow. The simulation results indicate that higher penetration rates of CAVs and CHVs contribute to the stability of complex mixed traffic flow, with CHVs exhibiting a more significant improvement effect than CAVs. Furthermore, considering information from multiple preceding and following vehicles has a greater impact on improving the stability and safety of complex mixed traffic flow compared to only considering information from immediately adjacent vehicles. Specifically, considering information from the two immediately preceding and following vehicles yields the best improvement in the stability and safety of complex mixed traffic flow.

In order to address the issue of insufficient safety assessment of autonomous vehicles compared to human driving, this study presents a driver emergency braking model that takes into account various risk scenarios. Firstly, through an analysis of environmental vehicle behavior, positioning, and traffic accident databases, three risk scenarios are identified: emergency braking of the front vehicle in the same lane, emergency lane change of the front vehicle in the adjacent lane, and encountering a stationary front vehicle after emergency lane change of the front vehicle in the same lane. Secondly, the study summarizes three key parameters of the braking model: maximum deceleration, braking efficiency improvement time, and decision response time, along with their extraction methods. These parameters were calibrated using both real vehicle testing and driving simulator experiments. The study defines the boundary between collision and non-collision in the risk scenario through the generalization of risk scenario variables and constructs the driver emergency braking model. Real vehicle testing of the typical scenarios derived from the generalization result shows that the proposed model has a 77.8% similarity to human driving capability. The parameters collected in this study can serve as a reference for future research on other driving models, and the constructed model can provide support for the assessment of the access and safety of autonomous vehicles.

Responding to demand of last-mile delivery programs can significantly improve customer satisfaction. Identifying and extracting characteristics of stopping points made by last-mile express tricycle deliveries is fundamental to analyze spatial-temporal distribution patterns and dynamic demand. This paper proposes a stopping point identification method by combining Point of Interest (POI) data with stopping time rules. The POI information and instantaneous speeds are utilized to screen express tricycle trajectory data. The stopping time threshold is introduced as secondary filtering criteria. The neighboring aggregation points are merged to create a complete set of stopping points. The accuracy of the identification results is verified through manual verification, and the entropy rate of the stop chain is calculated using the entropy rate method to quantitatively evaluate the accuracy of different identification methods. Taking trajectory data of express tricycles from Shun Feng Express courier outlets in Suzhou city as the empirical object, this paper compares the proposed method with the commonly used the density-based spatial clustering algorithm of applications with noise to identify stopping points of cargo trucks. The results show that the DBSCAN algorithm is prone to misidentifying traffic signal waiting as a delivery stop, while the proposed method effectively avoids this issue, achieving both precision and recall rates of up to 98%. Furthermore, the application of the entropy rate method further validates the effectiveness of the proposed method in terms of accuracy. On this basis, by expanding the research scope and identifying distribution stopping points, this paper analyzes the travel chains and spatial-temporal distribution characteristics of express tricycles. The results indicate that the number of delivery vehicles during the peak period around 8:00 am is significantly higher than that during the peak period around 4:00 pm. Residential areas are hotspots for distribution, with the highest concentration of vehicles, the longest travel distances, and the longest working hours. Hotel deliveries, on the other hand, exhibit shorter stopping times. Additionally, the spatial distribution of stopping points also reveals the delivery conditions to the remote locations.

Considering the unbalanced passenger flow distribution between parallel high-speed trains, this paper proposes a flexible ticketing method that considers passenger travel characteristics to alleviate pressure on benchmark train transport and to enhance overall ticket revenue and seat utilization rates. Flexible ticket is a new form of ticket system with lower prices and only a set of alternative train numbers are determined when sold, which has specific train numbers and seats being redeemed at a certain time before train departure. The departure time preference of passenger and the time value of flexible ticket delays are integrated into the passenger's choice of the generalized cost function. Based on the demand elasticity function, this paper introduces a passenger selection behavior model to evaluate the passenger flow allocation rate of each train. The model selects certain trains to offer flexible tickets on specific origin destination (OD) routes and develops an activation and cash model aimed at maximizing ticket revenue. The ticket allocation between non-alternative and alternative trains on each OD route is determined at various percentage stages as the decision variable. The model constrains include seat availability, train service section limitations and flexible ticket activation and redemption, which are established and resolved using Cardinal Optimizer (COPT). This paper used several trains on the Beijing-Shanghai high-speed railway to verify the model feasibility and effectiveness. The results indicate that the proposed method can quantify the ticket sales process from the supply and demand matching, which is not affected by spatiotemporal factors and passenger arrival probability. The method has high adaptability in the activation and redemption problem of flexible tickets. The discount of flexible tickets, the proportion coefficient of flexible tickets in the same percentage stage, the activation percentage stage, are key factors affecting the final return of passenger tickets. Properly increasing the activation and redemption frequency of flexible tickets can further improve ticket revenue. Under the condition of single point cashing activation, the maximum revenue increase is 4.84%, while under the condition of double point cashing activation, the maximum revenue increase was 5.38%.

To organize the overnight train of high-speed railway express, the interplay among the high-speed rail express demand, train operation and maintenance window setting should be addressed. Oriented by the demand of high speed rail express, this study collaboratively optimizes the express service plan to determine train timetable, express demand assignment and maintenance window setting. Under the mode of segmented rectangular maintenance window, it aims to minimize stranded express demand while ensuring express delivery timelines. The constraints of train operation, express delivery, and maintenance window setting are comprehensively considered. Then, a mixed integer nonlinear programming model is formulate for the high-speed rail express service plan. In order to reduce the difficulty of solving the model, a set of auxiliary variables are introduced to linearize the nonlinear objective function. The commercial optimization software GAMS with CPLEX solver is used to solve the model. Taking the Nanjing Hangzhou high-speed railway as a case study, a comprehensive overnight express service plan is obtained after 51 minutes of calculation based on parameter sensitivity analysis. The result illustrates that all 4000 planned shipments were served by the overnight express train, obtaining the optimal matching between high-speed rail express demand and trains. All shipments were delivered to their destination before 7:00 AM the following day, achieving the "next-morning arrival" rapid transit service. Additionally, the designated maintenance windows met the comprehensive maintenance requirements of the rail line. This study provides a decision-making basis for the high-speed express dispatching and train organization and provides a certain reference for the railway department to organize the future high-speed express night trains operation.

To reduce the impact of line interruptions on railway passenger transportation, this paper proposes an optimization method for the connection scheme when there is interruption of the train operations. The method is based on the complementarity between various modes of the comprehensive transportation network. The cost and passenger delay time of the connection scheme are analyzed from the changes in passenger train operation and connection scheme of comprehensive transportation transfer plan. Considering the factors of turnaround and arrival/departure capacity of transfer stations, the capacity of receiving vehicles, and the relationship between the station and vehicle stops, this paper develops a bi-objective programming model for passenger train unidirectional connection scheme optimization with the objectives of minimizing passenger travel time delay and the cost of the connection scheme. The algorithm is divided into multiple time periods based on the intensity of train arrival and departure time. The remaining turning capacity is analyzed to reduce the search space of the algorithm, and the feasible candidate set of train stations is established. A time-segmented Tabu Search algorithm is designed to solve the model. The results from the numerical example indicate that when a line interruption occurs, the connection scheme should choose a station with a closer connection distance. The average cost and total delay of transportation can be reduced by 45.3% and 21.2% compared to stop at the original arrival station. Therefore, using stations that no longer provide passenger train services for connection can reduce the impact of the line interruption. The proposed method can provide references for the decision makings under the line interruption conditions.

To address the path optimization problem in container road-rail intermodal transportation while considering hub delays, triangular fuzzy numbers are employed to characterize the uncertainty associated with such delays. This uncertainty takes into account the dual uncertainty in the number of containers awaiting transshipment and the remaining rail capacity in transport, as well as the limitations of hub transshipment capacity and rail departure requirements on hub delays. Therefore, the resulting delay time, along with the associated carbon emissions and storage costs, is quantified from the perspectives of waiting for transshipment and waiting for departure schedules. A multi objective path optimization model is then developed to minimize the total cost and total carbon emissions associated with the path scheme. The model is de-fuzzified using the expected value method and fuzzy chance-constrained programming. A self-adaptive fast nondominated sorting genetic algorithm (ANSGA-II) is designed to solve the model. Additionally, the simulation methods are used to assess the reliability of the path scheme and to identify the optimal combination of confidence levels. The case study was conducted on a specific road-rail intermodal transportation network, and the results indicate that, compared to the NSGA-II algorithm, the proposed method exhibits a faster convergence speed. The total cost and total carbon emissions were improved by 2.03% and 5.87%, respectively, and the reliability of the frontier solutions was greater than 0.95. Further sensitivity analysis indicates that when the number of containers awaiting transfer reaches a specific threshold, the preferred mode of transport tends to shift towards single road transport. This adjustment aims to minimize hub delays caused by waiting for transfer. At the same time, hub delays also have different effects on the path selection of goods with different time sensitivity.

To accurately describe the following behavior of air traffic flow, particularly considering the asymmetry of air traffic acceleration and deceleration, an air traffic flow following model based on an improved intelligent driver model (IDM) is proposed. Firstly, using actual flight trajectory data from the Chengdu terminal area, the following trajectory data is selected. The genetic algorithm is used to calibrate the parameters of the different classical following models, ultimately selecting the IDM model due to its superior fitting accuracy. Next, the improved IDM following models specifically for acceleration and deceleration scenarios are established considering factors such as acceleration, inter-vehicle spacing, average speed differences, and aircraft types. To assess the stability of the improved model, the Nyquist plot is used to analyze the local stability and asymptotic stability. Simulation experiments are conducted on the improved following model for acceleration and deceleration scenarios. The results show that the stable range of the improved IDM model is increased. For the acceleration and deceleration scenarios, the stability time of the improved IDM model is only 13.9% and 22.2% of the original IDM model, respectively. The model shows strong stability for different types of leading aircraft, and the leading aircraft of heavy requires the longest stable time, which is consistent with the practical observation that heavy aircraft generate the most severe wake turbulence in actual operations.

To accurately understand the influence of building distribution on station passenger flow, building census data based on building units was used instead of the POI (Point of Interest) data from existing studies. This approach considers various potential influencing factors such as building diversity, the number of transfer lines, the number of surrounding bus stops, and surrounding road mileage. After eliminating factors like road network density that show no significant correlation through preliminary regression analysis, a Geographically Weighted Regression (GWR) model that considers spatial heterogeneity was established to study the relationship between the total number of buildings and rail station passenger flow. Subsequently, by controlling factors such as the number of transfer lines and surrounding bus stops that affect station passenger flow, a Spatial Lag Model (SLM) that requires a smaller sample size was employed to analyze the impact of different types of buildings on rail station passenger flow. The study results indicate a positive correlation between the total number of buildings and station passenger flow, with correlation coefficients of 0.015, 0.007, and 0.004 at ranges of 500 meters, 800 meters, and 1000 meters, respectively. The correlation between building diversity and station passenger flow is not significant. Among various types of buildings, industrial and mixed use buildings do not significantly impact rail transit passenger flow, whereas other buildings show a positive correlation with station passenger flow. The correlation rank from highest to lowest is private residences, office, commercial, and residential buildings. As the distance from the station increases, the correlation of commercial buildings decreases most rapidly, followed by private residences and office buildings. Beyond 800 meters, the correlation between commercial buildings and station passenger flow is not significant.

To investigate the impact of different audio-visual warning patterns in vehicle-mounted intelligent early warning systems on driver recognition response in hazardous driving scenarios, this study conducted audio-visual bimodal warning experiments using a driving simulator platform, focusing on assessing the effectiveness of different audio-visual warning patterns under varying noise environments. The experimental design incorporated three factors: two noise environments (high and low) crossed with three visual warning patterns (0, 1, 2 Hz flashing) and three auditory warning patterns (non-speech, male speech, and female speech). Participants' performance data and subjective evaluation scores were collected, and a significance analysis was conducted to evaluate the effects. The results indicate that 1 Hz and 2 Hz flashing are more aligned with drivers' cognitive states and more acceptable than 0 Hz no flashing, with 1 Hz flashing significantly reducing reaction time. Male and female voice alarms are easier to understand and accept than non-verbal alarms, significantly decreasing recognition reaction time and improving accuracy. In high noise environments, drivers' comprehension of alarm information decreases significantly compared to low-noise environments. Across both high and low noise environments, the "flashing-voice" warning pattern yields the best warning effect, with "1 Hz flashing-male voice" achieving superior recognition performance, the highest perceptual matching, and comfort scores among the tested "visual-audible" warning patterns.

In response to the high frequency of traffic accidents on mountainous roads passing through villages and towns, this paper proposes a driving risk identification method considering driving style and reveals the influence of driving style in risk assessment through sensitivity analysis. The method is verified with a typical mountainous road passing through villages and towns in Yunnan Province as an example. The vehicle trajectory data is collected through drones and a trajectory database is established for the analysis. Based on the theory of driving risk field and through correlation analysis of accident data, the impact of driving environment on vehicle driving in the village and town section is characterized. Furthermore, driving style factors are introduced to establish a driving risk field that considers driving style, achieving comprehensive consideration of multiple factors and quantification of driving risk. Based on the Sobol global sensitivity analysis method, the global sensitivity of key parameters of the model before and after considering driving style is analyzed, and the visualization of regional risks is realized. The results indicate that when selecting the mean velocity, mean acceleration, and variance of impact for driving style classification, the clustering effect is best when the K value is 4. The Sobol method effectively evaluates parameter sensitivity. When considering the driver behavior field, the overall sensitivity distribution is more uniform, and the model considers a wider range of factors. Among them, the standard deviation of speed and impact are the most significant parameters, and the overall sensitivity is respectively 0.41 and 0.34. The visualization of the risks generated by the joint action of multiple vehicles shows that the potential field range varies with the shape of the road, and the risk value is most significant at the intersection through villages and towns, with an overall speed reduction ratio of 17.39%. This will further enhance the role of risk field theory in the field of micro traffic risk assessment.

To investigate the trajectory characteristics of diverging vehicles on the tunnel-interchange section with short connection, this paper collected driving data from driving experiment conducted from Huangcao Interchange to Baojia Interchange of Baomao Expressway. Drivers' lane choice along the tunnel-interchange section with short connection were analyzed. Distributions were evaluated for the diverging sites, longitudinal and lateral distances of diversion, and average sideway rates. The study also analyzed the effects of the connection length between the tunnel and the interchange, driver factors, and developed the trajectory characteristics model. The results show that drivers have no obvious lane preference and are more likely to miss the mainline exit in the tunnel-interchange section with short connection. With a reduction in length of the connection, the proportion of lane change increases in the tunnel entrance while decreases in the tunnel exit, the number of vehicles approaching the diversion point increases, the longitudinal diverging length exhibits a decreasing trend while the lateral displacement and the average traverse rate show an increasing trend. When the length of the connection is no more than 170 meters, there is a significant difference in the diverging characteristics between the tunnel-interchange section with short connection and the common interchange exit. These indicate with a reduction in length of the connection, the difficulty of vehicle lane change and diversion increase, and operation risk increase. On the short connection, the female drivers and the drivers who are not familiar with the route tend to change lane late, and the longitudinal diverging length is short. The inexperienced drivers are prone to changing directly from the left lane to the taper of the interchange, resulting in large lateral displacement, which might associate with high risk in diverging.