Long and bulky cargo has the characteristics of large outline, overweight and high cost and cannot be disassembled during the transportation process. Multimodal transportation is becoming the first choice of long and bulky cargo transportation, the core of which is route decision problem. In this paper, an energy consumption factor is introduced, and calculation formulas of carbon emissions during the transportation and reconstruction and reloading process at the node for long and bulky cargo multimodal transportation are all proposed. Then, taking into consideration the following factors, i.e., loading outline, gauge, bridge bearing capacity and reloading capacity at the nodes, and road reconstruction, a multimodal transportation route optimization model for long and bulky cargo with carbon emissions is proposed with the objective of minimizing multimodal transportation cost and carbon emissions. In addition, an adaptive genetic algorithm with an elite retention strategy is designed for the multimodal route decision for long and bulky cargo considering carbon emissions. Numerical results show that, compared with the traditional genetic algorithm and the adaptive genetic algorithm, the objective value of the proposed method is 20% higher and its cost and carbon emissions are 12% and 22% lower, respectively. The route plan by the proposed method can consider transportation cost and carbon emissions simultaneously, which can provide support to solve the multimodal route decision problem for long and bulky cargo and also reduce the cost and increase the efficiency in logistics and achieve the "dual-carbon" target.

This paper proposes an equity-oriented planning method for the freight carbon tax, which is developed based on an integrated modelling approach to study its potential impacts on regional equity. The proposed method uses a bi-level programming model. The upper-level model uses the Dagum Gini coefficient to evaluate the equity of the freight carbon tax policy on industrial locations across regions. Based on this evaluation, it optimizes the freight carbon tax rates for each region to enhance their equity. The lower-level model simulates the interaction among regional socio economic activities, transportation, and environment by adapting an integrated land use-transportation model. Due to the complex interaction between decision variables and objective functions, this study proposes a Bayesian optimization method to solve the above model. Extending from the above integrated land-use and transportation model, the proposed model not only provides a comprehensive impact analysis of the freight carbon tax policies on freight emission and industrial locations, but also takes the equity of such policies onto regional multimodal transportation networks and industrial locations into policy formulation. Taking the Yangtze River Economic Belt as an example, the findings suggest that a freight carbon tax policy of 80 Yuan per ton could intensify the adverse effects on the utility of industrial locations. Specifically, the net difference between regions accounts for 74.63% of the total Gini coefficients. In contrast, the differentiated freight carbon tax policy formulated by the model can not only effectively reduce the net difference between regions under the same freight emission reduction target, but also decrease the total Gini coeffifrom 0.180 to 0.115, which balances the equity impact of the freight carbon tax policies on the location of manufacturing industries in different regions.cient

The market-oriented price reform of railway freight transportation is an effective measure to promote the development of multimodal transportation and express logistics. Considering the shipper's selection behavior and the carrier's different freight business objectives, this paper proposed a framework and method for market-oriented pricing of railway "white goods" transportation. First, a freight mode selection behavior model is constructed to describe the shipper's mode selection behavior. Second, the pricing decision model is constructed under different freight business objectives in combination with clearing measures for rail freight receipts. The quantitative relationship between the attribute values of transportation services, such as railway freight price and total transportation time, and the indexes, such as railway freight volume, surplus and clearing income, is established. And the changes of various indicators are calculated after the unit price of "white goods" freight settlement has decreased. Then the pricing strategy of railway freight transportation is analyzed. The results show that at the current level of photovoltaic freight rates, the railways have achieved higher transportation volume, but the transportation surplus is negative and cannot maintain the current level of freight rates in the long term. The excessive proportion of transportation costs at two terminals of the railway transportation chain will weaken the impact of railway freight rate fluctuations on the choice of shippers. Reducing the clearing unit price of "white goods" provides more space for railway transportation to lower freight rates, which can increase the railway's share, surplus and clearing income, and stimulate the enthusiasm of railway carrier enterprises to explore the railway "white goods" transportation market.

To solve the problem of the actuator delay and uncertainties which may cause platoon instability or even destabilization, this paper proposes a robust model predictive control approach for vehicle trajectory optimization on urban roads. A third-order vehicle dynamics model was developed to optimize ride comfort, safety, platoon stability, fuel efficiency, and traffic delay. The behaviors of the red-light violations and the unsafe inter-vehicle distances were penalized, and the speed and acceleration were bounded. The signal changes were treated as system feedback. The proposed vehicle trajectory controller aims to improve the operational efficiency of controlled vehicles. The vehicle trajectory controller was formulated as a Min-Max model predictive control problem to enhance platoon stability by determining the control inputs in the worst case of actuator delays and uncertainties. Then, the iterative Pontryagin's maximum principle was used to solve the control problem, which discretized the control problem and divided the uncertain parameters into multiple intervals. To improve the computational efficiency, the proposed solution approach identified the worst case, iteratively computed the state variables forward in time, and solved the costate variables backward in time. The numerical simulation results demonstrate that the proposed controller performs well on the lane sections with and without signal controllers. The robust model predictive control approach can effectively response to random actuator delays and external vehicle disturbances, such as signal changes, abrupt speed changes, and small trajectory deviations caused by human drivers. The proposed robust Min-Max model predictive controller (MM-MPC) manifests better stability and superiority than the normal MPC controller in riding comfort (improved by 75.7%) and fuel consumption (reduced by 18.4%).

Connected vehicle (CV) has been proven to effectively improve traffic safety under fog weather conditions in microscopic driving behavior analysis. A microscopic car-following model is important for simulating the trajectory of CV in fog weather. Based on the traffic information perception mode and car-following behavior characteristics of CV in fog weather, this paper proposes a fog-related intelligent driver model of connected vehicle (FIDMCV) considering factors such as time headway, weighting, and compliance, based on the fog-related intelligent driver model. To evaluate the effectiveness of the FIDMCV model and assess the traffic impact of CV in fog weather, the cumulative reciprocal of Time-to-collision (1/TTC) and throughput were selected as analysis indicators, and numerical simulation scenarios with different CV penetration rates and decelerations of the leading vehicle were established. Before conducting numerical simulations, sensitivity analyses were performed on key parameters of time headway and compliance. The simulation results show that with the increase in the penetration rate of CV, mixed traffic flow more effectively improved traffic safety in fog weather. However, it also led to an increase in car-following distances of vehicles, thereby reducing road throughput and decreasing traffic efficiency. The proportion of reduction in cumulative 1/TTC values for CV in a high risk scenario (deceleration of 6 m⋅s²) is 14.3%, and in medium-low risk scenarios (decelerations of 4 m ⋅ s² and 2 m ⋅ s²) is 5.6% and 6.3%, respectively, indicating that the improvement of traffic safety for CV is more significant in the high risk scenario. The proposed FIDMCV model can effectively reflect the traffic safety improvement effect and car-following distance increase characteristics of CV in fog weather conditions, and can be used as a microscopic simulation tool for CV.

The centralized nature and flexibility of electric vehicle (EV) ride-hailing fleets offer opportunities for vehicles to provide emergency and demand-response services to the grid during peak load periods, when combined with Vehicle-to-Grid (V2G) technology. This study investigates the flexibility of EV ride-hailing fleets participating in V2G systems and aims to make dynamic decisions on vehicle-trip assignment, empty vehicle relocation, and charging/ discharging schedules. First, a time-space-energy three-dimensional network is constructed to depict the vehicle scheduling problem. Then, the rolling horizon optimization model is used to maximize the expected benefits of the fleet. Additionally, the dynamic scheduling decisions of the fleet are obtained by defining feasible arcs. A case study is conducted in Jiading, Shanghai. The results indicate that the proposed strategy for EV ride-hailing fleets can effectively respond to travel requests, balance future travel demand and supply through empty vehicle relocation, and dispatch idle vehicles for discharging. During periods of grid demand response, 10.3% of idle vehicles can be dispatched for discharging, with an average revenue of 104.8 yuan per hour per vehicle. The proposed method helps reduce vehicle idle rates, increase vehicle revenue, and address the issue of the gradually saturated transportation service market.

For dynamic scenarios involving interaction among multiple vehicles, intelligent vehicles should be able to predict the future trajectories of surrounding vehicles for safe and efficient driving. This paper proposes a trajectory prediction method that considers dynamic interactions among vehicles. First, based on the historical trajectory information of the target and surrounding vehicles, a dynamic spatio-temporal correlation graph is constructed as the input for the interaction feature extraction module. The graph attention mechanism is then used to capture the temporally varying interaction feature parameters. Second, the historical temporal information of the target vehicle is fused with the variable interaction feature parameters. A context vector is obtained by an LSTM encoder embedded with a temporal attention mechanism, followed by using the LSTM decoder to output the future trajectory of the target vehicle. Finally, the proposed model is trained and validated on the CitySim dataset, and transfer experiments are conducted using the CQSkyEye dataset. The results show that the model achieves an RMSE of 0.82 m in a 5 s prediction horizon, demonstrating a 15% improvement in accuracy compared to other popular models. The model also demonstrates the ability to make predictions with less than 2 s lead time. In terms of transferability, the proposed model outperforms others with an RMSE of 6.43 m in the 5 s prediction horizon after adjusting the distance threshold parameter for graph construction, showing an improvement of over 48% in transfer prediction capability.

This paper proposes an adaptive control model for bus priority at an intersection based on speed guidance to minimize the average waiting time per passenger at the intersection, This method further satisfies the demand for real time priority for buses at the intersection with different volume scenarios. The relationship between bus speed and signal timing parameters is established, and the simultaneous optimization of the intersection signal timing scheme and bus speed guidance scheme is realized by taking into account the real-time volume of the intersection approach and the real- time operation status of buses. To lower the computational complexity of the model and increase its speed of solution, the phase sequence of the intersection is represented by the 0-1 decision variable. The effectiveness of the model is confirmed by simulation experiments. When compared to the comparison schemes for the three different traffic volume scenarios, the proposed model can reduce average bus delays by at least 17.86%, 12.04%, and 8.81%, average bus stops by 21.82%, 17.86%, and 17.74%, and average passenger waiting times by 24.56%, 8.03%, and 3.38%. Furthermore, the model's solution time for different traffic scenarios is less than 0.01 seconds, which is a significant reduction in computing time when compared to the traditional models and further satisfies the real-time bus priority calculation efficiency requirements. The proposed model can meet the requirement for model optimization rate in the networked environment while simultaneously achieving bus priority and lowering the average waiting time per passenger at intersections without compromising the overall intersection and social vehicle efficiency.

In response to the problem of low driver recognition accuracy caused by the low resolution of monitoring images and the weakening of vehicle motion trajectory features in tunnel environments, this paper proposes a driver recognition method (Multi-scale Convolutional Neural Network with Multi Attention) that integrates convolution with multi-head attention mechanism. The collaborative coupling relationship of multi-source information of human-vehicle road-environment during the driving process was utilized to improve recognition accuracy. First, real vehicle driving experiments are designed and conducted to establish a multi-source driving database for tunnel sections and design feature sets. Second, a driver recognition model framework is built. This framework learns local fluctuations in the driving process through a multi-scale convolutional neural network and captures the long-term dependency of driving time series through parallel multi-head self-attention layer structures, effectively integrating local and global information to enhance the driver recognition effect in tunnel scenarios. The results show that compared with other advanced algorithms, the proposed model achieves an accuracy of 99.07% and a macro F1 score of 99.03% in driver identification tasks, fully demonstrating the effectiveness of the proposed method. In addition, through the feature contribution evaluation method, the importance of features in driver identification tasks in tunnel scenarios is explored in depth. It is found that compared with historical vehicle motion data, driver psychological, physiological, and visual features show higher contribution. The research results can provide support for the application of multi-source data in tunnel scenarios and provide technical support for tunnel safety supervision.

Aiming at the congestion problem in the expressway entrance ramp area, this paper proposes an entrance ramp control method based on traffic density prediction. By analyzing the traffic characteristics of the expressway ramp area, this study defines the slow traffic flow state, use the classic macro traffic flow model to make statistics on the historical density of the expressway mainline, and perform cluster analysis based on speed and density parameters to quantify the traffic parameter characteristics of the slow travel state. Based on the diurnal variation characteristics of urban traffic flow, a time series model is used to make short-term predictions of traffic density, and slow traffic conditions are identified based on the prediction results. In view of the slow traffic state, a collaborative control model is proposed for the traffic density of the expressway mainline and the queue length of the entrance ramp. By controlling the entrance ramp, the density downstream of the mainline is controlled not to exceed the critical density, and the PSO PID algorithm is used to solve the control model. Taking the Hongdu Avenue Expressway in Nanchang City as an example, a simulation analysis was conducted on the evening peak slowdown period, and a comparative analysis was conducted on the classic ramp control scheme. The results show that compared with the ALINEA model, the proposed model increases the ramp adjustment rate by 11.0%, and reduces the average queue length and average delay of the ramp respectively by 14.8% and 13.5%.

In order to accurately reproduce the interaction behavior of bicycles to meet the needs of autonomous driving simulation testing, a Position Reward Augmented Generative Adversarial Imitation Learning (PRA-GAIL) method is proposed. In urban roads, since the disturbance behavior is mainly generated by electric bicycles, electric bicycles are selected as the research object. In the constructed simulation environment, Generative Adversarial Imitation Learning (GAIL) is used to make the simulated trajectories approximate the real trajectories, while Position Reward and Lagrangian Constraint methods are added to solve the homogenization and uncontrollable behaviors of existing simulation methods. In the test set validation, the average displacement error of the GAIL and PRA-GAIL methods decreased by 61.7% and 65.8% , respectively, compared to the behavioral cloning method. In the behavioral performance validation, the KL divergence of acceleration distributions between simulation and reality was significantly reduced in PRA-GAIL compared to GAIL, and the percentage error of overtaking and illegal lane changing behaviors decreased by 7.2% and 20.2%, respectively. Using the Lagrangian method to add constraints resulted in a 75.8% reduction in the number of agents with risky behavior compared to commonly used reward augmentation methods. In trajectory validation, in the simulation environment, the average displacement error of PRAGAIL is reduced by 17.5% compared to GAIL. The resulting model realistically reproduces the overtaking maneuver space of cyclists. The results show that the method adopted in this paper is suitable for bicycle behavior simulation, the proposed modifications effectively enhance the simulation performance, and the obtained simulation model accurately reproduces the disturbance behavior of bicycles on road segments, which can be applied to automated vehicle simulation tests.

To investigate the changing characteristics of driving psychological load in the tunnel portal sections of a two-lane highway in a mountainous area, this paper conducted a real vehicle test with 27 drivers as the subjects in nine tunnels in Chongqing. The driving data of drivers, vehicles and driving environments were collected, and the heart rate growth of the drivers was used to characterize the psychological load of driving. The correlation between drivers' driving psychological load and luminance and speed in the tunnel portal section was investigated through data analysis. The results showed that the psychological load of driving in the entrance section of a two-lane highway tunnel in a mountainous area was greater than that in the exit section. When the ambient luminance is less than 500 cd·m2, the driving psychological load is more sensitive to the luminance change rate, and then the formula for the luminance change rate is proposed. When the luminance change rate is 60%, the psychological impacts of the black hole and white hole effects on drivers tend to be the same. In the tunnel portal section, speed has an amplifying effect on driving psychological load. Based on the driving psychological load value, this study also defined the driving comfort interval of the tunnel portal section, proposed the safety thresholds of luminance change rate and speed based on the driving comfort demand, and developed the luminance curves in consideration of the luminance change rate and the clearance height of the tunnel portal.

To describe the specific forms of dangerous driving behavior and quantify the overall driving risk of drivers on ice and snow pavement, driving operation data and vehicle motion data were collected through simulated driving experiments. A dangerous driving behavior spectrum was constructed, which included five behaviors: sharp changes in direction, sharp changes in speed, speeding, lateral swaying, and occupying adjoining lanes. The threshold values of dangerous driving behavior under different pavement adhesion coefficients were determined based on specific experimental conditions and the quartile difference method. Then, the characteristic value of the dangerous driving behavior spectrum was calculated by weighted averaging. The characteristic value of the dangerous driving behavior spectrum was used as the driving risk evaluation index. The K-means clustering algorithm was used to divide the driving risk state into four categories, and the Random Forest model, metaheuristic optimization algorithm, and Support Vector Machine (SVM) were combined to establish a model for identifying driving risks on ice and snow pavement. The results show that speeding has the greatest impact on driving safety on ice and snow pavement with a weight of 0.285; the characteristic values of the dangerous driving behavior spectrum are mainly distributed in the interval [0.00, 0.20], and the maximum value appears under the pavement adhesion coefficient of 0.2; under the conditions of pavement adhesion coefficients of 0.2 and 0.4, the severe driving risk increases significantly, accounting for 4.85 times and 2.49 times of normal pavement, respectively. The recognition accuracy of the driving risk recognition model obtained by combining the Osprey Optimization Algorithm (OOA) with SVM is 93.96%, outperforming the Coati Optimization Algorithm (COA) and the Grey Wolf Optimizer Algorithm (GWOA). The research results provide insight into the understand the differences between driving on ice and snow pavement and normal pavement, and provide a basis for enhancing driving safety.

To reduce the high energy consumption of battery electric buses (EBs) when crossing the signalized intersections, this paper proposes a rule-based two-stage driving strategy, which accounts for reasonable strategy scenarios and different acceleration and deceleration variation concave-convex without sacrificing passenger comfort and passing time. In the first stage, four driving strategies are proposed for EB crossing the intersections, including constant speed strategy, acceleration strategy, deceleration strategy, and brake stopping strategy. This process considers vehicle operating status and the signal phase and timing (SPaT) information of the intersection obtained via vehicle-to infrastructure (V2I) technology. In the second stage, the energy consumption estimation model based on the XGBoost algorithm and the curve model based on the acceleration and deceleration characteristic parameters are used to optimize the speed profile of the constant speed process, acceleration process, and deceleration process in the selected driving strategy. The effectiveness of the four driving strategies was analyzed by the VISSIM simulation. The results show that the proposed strategy can effectively reduce energy consumption by 29.8%~34.2% in four driving scenarios. Further analysis revealed that the optimal speed profile has concave-convex properties both above and below the economic speed, therefore the concave-convex laws of the most energy-saving speed profiles with different speed adjustment ranges are summarized through comparative experiments, which can Provide easier-to-operate driving guidance for energy-saving driving of EBs.

In order to improve the operating efficiency and reduce the operating cost of electric bus systems, this paper proposes an electric bus scheduling optimization method that considers opportunity charging and travel time reliability. Firstly, based on the regional scheduling scenario, a strategy of equipping fast-charging piles at the beginning and end stations of the lines and utilizing the succession time for opportunity charging is proposed. Then, considering the stochastic fluctuation of travel time, the reserved travel time characterizing the specific reliability is used as the model input to generate the scheduling scheme, and the departure delay cost is incorporated into the objective function. Considering the overall benefit from the planning to operation stages, a regional multi-model electric bus scheduling optimization model aiming at the minimum total cost was constructed, and an adaptive large-neighborhood search algorithm was designed to solve the model. Finally, four bus lines in the Daxing District of Beijing are taken as examples to verify the effectiveness of the model and the algorithm. The results show that compared with the traditional single-route single-vehicle type scheduling scheme, the optimal scheme based on the proposed method can reduce the daily average cost of bus companies by 37.93%, and the average departure delay time of each vehicle is reduced by 5.63 minutes, which indicates that the proposed method can effectively reduce the cost of enterprises and improve the reliability of public transportation system. Compared with the regional multi-vehicle operation model without considering the opportunity charging strategy and travel time reliability, the optimal scheme in this paper can reduce the total cost by 28.67%. In addition, through the sensitivity analysis, it is suggested that the bus companies should configure the fast-charging resources with 240 kW charging power and prepare the electric bus scheduling scheme with 90% travel time reliability.

Traditional fixed-capacity buses struggle to meet the varying demand distribution on bus routes. To tackle this challenge, modular buses are introduced, allowing for dynamic adjustments in platoon capacity through joining and detaching, thus better accommodating spatial demand variations. An optimization model is developed to describe the operational scheme of modular bus routes, based on the reconstruction of spatiotemporal graphs. The formulated model, a Mixed Integer-Nonlinear Program (MINLP) model, includes decision variables such as platoon schemes and modular bus unit schemes. To facilitate the model solution, time discretization is applied, which transforms the MINLP model into a Mixed Integer-Linear Program (MILP). A case study is performed using real bus routes and passenger demand data from Chengdu, China. Experimental results demonstrate that the use of modular buses reduces passenger costs by 11.44% and operating costs by 31.35% compared to traditional fixed-capacity buses, resulting in an overall decrease of 20.32% in total system costs. Sensitivity analysis experiments examine the effect of system supply and demand changes on system costs.

Pure electric bus has become an important option for the electric transformation of vehicles due to its low carbon, energy-saving, and environmental protection characteristics. However, pure electric buses still face challenges such as performance degradation under low-temperature conditions and reduced mileage due to battery aging in actual operation. The mixed use of fuel buses and pure electric buses in operation helps to improve the performance degradation of pure electric buses in specific scenarios, and to enhance the efficiency and service quality of bus operation. This paper proposes a segmented optimization model for bus timetables considering the dynamic operation characteristics of buses. With the optimized frequency as input, a bus fleet scheduling planning compilation model is developed under mixed bus operation conditions. An improved genetic algorithm is designed to solve the model. Taking the bus routes in Beijing as an example, the case studies were conducted in different typical operational scenarios such as single-line operation, remote charging, and regional centralized scheduling to verify the applicability and optimization effect of the model under differentiated operational scenarios. The results indicate that compared to local charging scenarios, operational costs increased by 5.15% and the number of operating vehicles increased by 5.88% in remote charging scenarios. In the regional centralized scheduling scenario where multiple routes are jointly scheduled, operational costs decreased by 4.68% compared to single-line operation scenarios. Under the condition of given bus types proportion threshold, the effectiveness of mixed vehicle operation surpasses single vehicle type operation, effectively reducing operational costs and carbon emissions. This study provides a support for public transport enterprises to create scientific and flexible electric bus operation scheduling schemes based on different operation scenarios.

This paper proposes an optimization model for electric bus scheduling and charging scheduling with the consideration of time-of-use electricity pricing policy and multiple vehicle types, aiming to minimize the total operation cost of electric bus system. The practical operational constraints of bus schedule chain formulation, charging time window, and limited number of chargers are considered in the model. An adaptive large neighborhood search (ALNS) algorithm is proposed to solve the bus schedule optimization problem. This algorithm incorporates diverse destruction and repair operators tailored to the characteristics of the problem, such as the trip-to-vehicle allocation and the feasibility of the bus schedule chain under multiple vehicle types. For the feasible bus schedule chain combinations generated by ALNS, the charging schedule optimization subproblem under time-of-use electricity price is constructed and mapped into a dedicated network. An algorithm based on the minimum-cost-flow is designed to solve for the charging duration, which leads to an optimal decision on charging start time. The model and algorithm are validated using three bus routes in Beijing. The results show that compared with the current situation, the fleet size is reduced from 30 to 24 vehicles, resulting in a decrease in electricity cost and total operation cost by 25.84% and 20.63%, respectively. Comparative experiments are conducted to explore the impact of different weights of repair indicators and combinations of vehicle types on the optimization results.

This study investigates the impact of the built environment on the demand for dockless bike-sharing (DBS) and integrated metro use. A number of 120,000 DBS trip records were utilized, and spatial confidence ellipse technology was employed to illustrate the clustering characteristics of DBS near metro stations. Subsequently, a quantitative method for delineating bicycle-metro catchment areas was developed, through which the built environment surrounding metro stations was evaluated through five dimensions: density, transportation facilities, land use, destination accessibility, and metro ridership. Finally, a gradient boosting decision trees (GBDT) model is employed to map the complex and non-linear interactions between the built environment and the necessity for integrated use modalities. The results indicated that metro ridership and workplace locations emerged as significant factors influencing the integrated use, exhibiting a distinct threshold effect. An increase in commercial activities initially elevates the integrated travel demand, but excessive density subsequently triggers adverse effects due to traffic congestion. An uptick in bus stop density indicates a competitive dynamic between shared bikes and public transit, underscoring the intricate interactions within urban transportation systems. Furthermore, the nonlinear effects of land use diversity and population density underscore the profound relationship between urban planning and residents' commuting behaviors.

Passenger travel in rail transit networks is affected by changes in network structure and operating conditions, and individual travel preferences respond differently to these changes. In order to analyze the impact of a planned shutdown of a suburban rail line on the travel transfer of frequent passengers, a passenger travel feature characterization method considering transfer types and transfer ratios was proposed, and the passenger's feature temporal (F-T) matrix was generated by combining time period attributes. The similarity between F-T matrices was calculated by an improved Euclidean distance to achieve the similarity measurement of F-T matrices. A two-step clustering method of K-Means clustering and hierarchical clustering (KMHC) based on the similarity matrix was proposed to partition the affected passenger groups, and the factors affecting passenger transfer were analyzed. The Kunshan section of Shanghai Rail Transit Line 11 during COVID-19 was taken as an example to verify the method. The research results show that after the shutdown of the Kunshan section, there are five main groups of travel transfer impacts of frequent passengers, accounting for 94.4% of the total number of frequent passengers. The transfer distance, commuting time and travel frequency of the affected groups are obviously different, which are important factors influencing the travel choices of frequent passengers after the section shutdown. The method can serve as a reference for other planned shutdown scenarios, and can also provide support for predicting changes in network passenger flow, and optimizing driving and passenger transportation organization plans after the section shutdown.

This paper studied the inventory routing and pricing problem for e-commerce companies operating in the omnichannel mode. Considering the uncertain demand factor of each front warehouse, a differentiated pricing strategy for goods in different selling channels was proposed. A mixed-integer nonlinear robust optimization model was constructed with the objective of maximizing the total profit. The e-commerce company's risk attitude on the demand uncertainty is set by the conservative coefficient. And then an adaptive simulated annealing particle swarm algorithm was designed to solve it. Two sets of examples were selected, including 10 and 20 front warehouses, to verify the applicability and effectiveness of the proposed model and algorithm. The results of experimental analyses show that differentiated pricing can increase the total profit of the e-commerce company by about 5% and 6%, respectively, compared with uniform pricing. The results of the sensitivity analysis indicate that, enhancing offline shopping experiences of online customers to increase the number of customers who buy online and pick-up in store, and organizing online marketing activities to increase the sensitivity of online customers to e-commerce promotion efforts, can bring higher profits to e-commerce companies. Controlling future market volatility risks and accurately predicting demand information to reduce e-commerce companies' conservative coefficients and maximum demand deviation coefficients can also increase total profits of e-commerce companies. The findings of the study can provide a reference for e-commerce companies to formulate inventory routing strategies for their front warehouses and goods pricing schemes for their selling channels.

To solve the time-dependent vehicle routing problem with simultaneous pickup-delivery and time windows (TDVRPSPDTW), this paper proposes a mathematical model with the sum of vehicle fixed cost, driver cost, fuel consumption and carbon emission cost as the optimization objective. Based on the traditional ant colony optimization, this paper introduces a hybrid ant colony optimization with adaptive large neighborhood search (ACO-ALNS). It uses heuristic initialization of pheromones, improves state transition rules, and uses local search strategies to improve solution quality. Benchmark problem instances and adapted TDVRPSPDTW instances are utilized for experimentation. The experimental results demonstrate the effectiveness of the proposed ACO-ALNS algorithm in solving the benchmark problem of TDVRPSPDTW. Compared to the simulated annealing and ant colony optimization with local search, the proposed algorithm improves the optimal value of total distribution cost by an average of 7.56% and 2.90%, respectively. In addition, the presented model incorporates multiple factors, resulting in an average reduction of 4.38% and 3.18% in total distribution costs compared to models that only consider carbon emissions or delivery time. This improvement can effectively enhance the economic benefits of logistics enterprises.

To improve the participation between control and airlines in collaborative decision-making and reduce flight delays, this paper proposes a multi-objective trajectory slot allocation model considering airline preferences to solve the problem of route resource allocation with multiple restricted zones. The model aims to meet capacity constraints, minimize flight delays, and minimize the associated costs of rerouting as efficiency and fairness objectives. The study applied the model to a domestic route instance data and used the Benders decomposition algorithm to analyze and solve the model. The results show that compared with traditional ground waiting methods, the total delay cost per minute of flights was significantly reduced by 6.2%. Due to the reduction of ground delay caused by diversion allocation, the total delay time of flights was reduced by 29.3%. This result emphasizes the importance of increasing rerouting (except for ground waiting) to avoid airspace restricted areas. In addition, the model allows airlines to flexibly choose original or alternative routes based on their own preferences, and analyzes the impact of airline preferences on individual and system delay levels. At last, the trade-off between efficiency and fairness in multi constrained resource allocation under different fairness schemes was discussed. The proposed Min-Max method can improve airspace operational efficiency by 5.4% and airline fairness by 70.8% compared to existing RBS (Ration-By-Schedule) methods. It can be seen that the proposed multi-objective collaborative route allocation optimization method is effective in solving the problem of collaborative route resource allocation, taking into account the fairness of various airlines while reducing the total delay cost.

Machine learning models have been extensively applied in exploring the interaction between the built environment and passenger flow. However, machine learning primarily considers global relationships and fails to capture spatial variations. To address this issue, this paper defined 11 built environment variables from the aspects of density, diversity, design, destination accessibility, availability, and network connectivity. The study proposed an integrated ensemble analysis model, SLightGBM, combining Light Gradient Boosting Machine (LightGBM) with Geographically Weighted Regression (GWR), to investigate the spatial heterogeneity and nonlinear impact of the built environment on the attractiveness of station coverage. The SLightGBM model was compared with the LightGBM, Ordinary Least Squares(OLS), and GWR to demonstrate its regression superiority. The results from Xi'an city indicate that: (1) The SLightGBM model showed better performance than other models, with R2 value of 0.68, MAE of 8379.16, and RMSE of 11797.19. (2) The factors of the built environment vary across spaces. The densities of the employment and bus stops are most important in central areas, whereas the density of restaurants is more prominent in the southern regions. (3) Higher employment and restaurants densities are positively correlated with the attractiveness of metro ridership, while the minimum transfer times are negatively correlated with the attractiveness of metro ridership, showing a strong combined effect. This study indicates the importance of understanding spatial differences and threshold effects of these factors in urban planning and public transport system improvement.

In order to deeply analyze the evolutionary characteristics of air traffic congestion situations, support air traffic control and operation of an air traffic route network, this paper conducted a study on congestion propagation of the air traffic route network. First, segment traffic flow, segment traffic density, and segment traffic convergence were selected as segment congestion identification indicators, and an FCM segment traffic state evaluation model was established. Second, a congestion propagation model based on transfer entropy theory was proposed, and the correlation between congestion and congestion propagation was analyzed. Finally, a method for identifying key routes based on propagation index and significant area index was developed. Real-measured data from the Guangzhou area control sector was used to validate the effectiveness of the proposed methods. The research results indicate that a congestion identification model that considers both macroscopic and microscopic features can effectively classify the traffic status of routes. Congestion propagation is strongly influenced by time and congestion levels. During nighttime periods, smooth traffic dominates in congestion propagation, with information dissemination being 30% higher than during daytime, and with a longer information validity period. Conversely, during daytime, heavily congested routes dominate congestion propagation, with information dissemination being approximately 1.2 times that of lightly congested routes, but with a shorter information validity period. Congestion propagation in the Guangzhou high altitude route network exhibits significant spatiotemporal variations, with congestion propagation peak periods mainly occurring around 12:00-14:00 and 18:00-20:00, slightly earlier than congestion peak periods. Key routes for congestion propagation include major east-west routes such as the Shanghai-Chongqing route. These patterns provide theoretical support for air traffic control units to implement control measures and optimize route structures.

To explore the mental workload characteristics of drivers on urban underground ring roads, a natural driving test was conducted, and physiological instruments were used to obtain drivers' electrocardiogram data in their natural driving state. A quantitative model of mental workload was constructed based on the factor analysis, using drivers' heart rate, heart rate growth rate, and heart rate variability indicators as key variables, to reveal the effects of underground ring road flat curve radius, the curve type, entrance and exit, and the proficiency level on drivers' mental workload. Results show that heart rate index and heart rate variability index are negatively and positively correlated with the radius of the underground ring road flat curve, respectively. The mental workload of drivers is negatively correlated with the radius of the flat curve. Under the same radius conditions, the mental workload of consecutive curve scenes is significantly higher than that of single curve scenes. Furthermore, the cumulative effect of consecutive curves exceeds that of single curve radius on drivers' mental workload. The mental workload of drivers at the entrance segment is 21.39% higher than that at the exit segment. The mental workload of non-proficient drivers is 30.75% higher than that of proficient drivers. It is recommended that sight-inducing facilities be added at the entrances and exits of urban underground ring roads and sections with curvature changes, and the flat curve radius should not be less than 30 m when the design speed is 20~40 km·h^-1.

Electroencephalogram (EGG) signals, with their high temporal resolution among other advantages, have become an essential tool for recognizing drivers' cognitive states and assessing driving performance. Previous research on brain electrical activity in the context of driving behavior has often been limited to abnormal driving states, such as fatigue detection and distracted driving, neglecting normal driving scenarios. This paper focuses on regular driving behaviors recognition. Through driving simulation experiments, this study synchronously collected driving and brain electrical activity data from drivers while they performed acceleration, deceleration, and turning maneuvers. A channel attention-separable convolutional neural network based on the squeeze-and-excitation module is constructed to recognize the driving behaviors and optimize channel selection across individuals' brain electrical signals. The results show that the proposed model achieved an accuracy of 82% in recognizing three types of driving behaviors while reducing the number of channels by 70% without compromising prediction accuracy. The effectiveness of the model was demonstrated through ablation experiments and comparisons with other baseline models. Analysis of the optimal channel combinations' scalp topography revealed that the frontal and occipital areas of the brain are most relevant to regular driving behaviors. The findings of this study provide a methodological basis to understand driving behavior from a cognitive perspective and for brain-like driving decision-making.

To analyze the advantages of hydrogen fuel cell heavy-duty vehicles (FCHDVs) compared to conventional energy heavy-duty vehicles under China's current energy structure, this study employed the life cycle assessment method and the GREET model developed by the Argonne National Laboratory. A controlled variable method was used to compare the full life cycle air pollutant emissions of FCHDVs and diesel heavy-duty vehicles under the coal-to hydrogen and renewable energy-to-hydrogen pathways. A sensitivity analysis was conducted on factors affecting vehicle emissions. Finally, a fleet replacement model analysis was performed using the heavy-duty vehicle fleet in Yulin City as a case study. Simulation results indicate that FCHDVs using the renewable energy-to-hydrogen pathway reduce various air pollutant emissions by an average of 68.74% compared to diesel heavy-duty vehicles. FCHDVs using the coal-to-hydrogen pathway emit lower levels of CO, NOx and CO2 gases on average than diesel heavy-duty vehicles. Hydrogen production from renewable energy generation reduces air pollutant emissions from hydrogen fuel cell heavy-duty vehicles by an average of 92.41% compared to ordinary electrolysis hydrogen production. Long-haul FCHDVs reduce air pollutant emissions by 12.57% compared to short-haul FCHDVs. Replacing diesel heavy-duty vehicles with FCHDVs is expected to reduce the overall fleet's air pollutant emissions by 13.77% by 2030. This paper analyzes the emission levels of air pollutants from FCHDVs in the context of road transportation and provides a basis for the subsequent promotion of FCHDVs.

Autonomous minibuses play an important role in the micro-circulation of public transportation. To achieve commercial application of autonomous driving technology, in addition to traditional vehicle performance tests, it is also necessary to evaluate their performance in complex scenarios. Due to the lack of test data, the limited applicability of evaluation models, and the subjectivity of evaluation methods, previous evaluations have shown significant biases. This study developed a comprehensive evaluation system for the performance of autonomous minibuses and conducted a comprehensive evaluation based on field test data using a TOPSIS model combined with game theory weighting. Four evaluation dimensions were selected: driving safety, ride comfort, vehicle intelligence, and vehicle efficiency, which were further divided into 12 objective evaluation indicators. First, data were collected through field tests in the operating scenarios of autonomous minibuses. Second, the game theory-based combined weighting method was used to combine the weights obtained from the analytic hierarchy process and the entropy weighting method. Finally, to verify the effectiveness of the model, the TOPSIS model was used to calculate the comprehensive evaluation values of three test routes with different complexity. The results show that in the performance evaluation of autonomous minibuses, the order of importance at the criterion level is vehicle intelligence, driving safety, driving comfort, and vehicle efficiency, while the sensitive indicators at the index level are autopilot state and average angular velocity. The results of the performance evaluation of autonomous minibuses on routes with different scenario complexities using the game theory based combined weighting TOPSIS model were consistent with the actual operating conditions, demonstrating the effectiveness of the method.