A novel passenger flow prediction model using deep learning methods

Currently, deep learning has been successfully applied in many fields and achieved amazing results. Meanwhile, big data has revolutionized the transportation industry over the past several years. These two hot topics have inspired us to reconsider the traditional issue of passenger flow prediction. As a special structure of deep neural network (DNN), an autoencoder can deeply and abstractly extract the nonlinear features embedded in the input without any labels. By exploiting its remarkable capabilities, a novel hourly passenger flow prediction model using deep learning methods is proposed in this paper. Temporal features including the day of a week, the hour of a day, and holidays, the scenario features including inbound and outbound, and tickets and cards, and the passenger flow features including the previous average passenger flow and real-time passenger flow, are defined as the input features. These features are combined and trained as different stacked autoencoders (SAE) in the first stage. Then, the pre-trained SAE are further used to initialize the supervised DNN with the real-time passenger flow as the label data in the second stage. The hybrid model (SAE-DNN) is applied and evaluated with a case study of passenger flow prediction for four bus rapid transit (BRT) stations of Xiamen in the third stage. The experimental results show that the proposed method has the capability to provide a more accurate and universal passenger flow prediction model for different BRT stations with different passenger flow profiles.     

Inferring patterns in the multi-week activity sequences of public transport users

The public transport networks of dense cities such as London serve passengers with widely different travel patterns. In line with the diverse lives of urban dwellers, activities and journeys are combined within days and across days in diverse sequences. From personalized customer information, to improved travel demand models, understanding this type of heterogeneity among transit users is relevant to a number of applications core to public transport agencies’ function. In this study, passenger heterogeneity is investigated based on a longitudinal representation of each user’s multi-week activity sequence derived from smart card data. We propose a methodology leveraging this representation to identify clusters of users with similar activity sequence structure. The methodology is applied to a large sample (n = 33,026) from London’s public transport network, in which each passenger is represented by a continuous 4-week activity sequence. The application reveals 11 clusters, each characterized by a distinct sequence structure. Socio-demographic information available for a small sample of users (n = 1973) is combined to smart card transactions to analyze associations between the identified patterns and demographic attributes including passenger age, occupation, household composition and income, and vehicle ownership. The analysis reveals that significant connections exist between the demographic attributes of users and activity patterns identified exclusively from fare transactions.     

Validating and improving public transport origin–destination estimation algorithm using smart card fare data

Smart card data are increasingly used for transit network planning, passengers’ behaviour analysis and network demand forecasting. Public transport origin–destination (O–D) estimation is a significant product of processing smart card data. In recent years, various O–D estimation methods using the trip-chaining approach have attracted much attention from both researchers and practitioners. However, the validity of these estimation methods has not been extensively investigated. This is mainly because these datasets usually lack data about passengers’ alighting, as passengers are often required to tap their smart cards only when boarding a public transport service. Thus, this paper has two main objectives. First, the paper reports on the implementation and validation of the existing O–D estimation method using the unique smart card dataset of the South-East Queensland public transport network which includes data on both boarding stops and alighting stops. Second, the paper improves the O–D estimation algorithm and empirically examines these improvements, relying on this unique dataset. The evaluation of the last destination assumption of the trip-chaining method shows a significant negative impact on the matching results of the differences between actual boarding/alighting times and the public transport schedules. The proposed changes to the algorithm improve the average distance between the actual and estimated alighting stops, as this distance is reduced from 806 m using the original algorithm to 530 m after applying the suggested improvements.     

Carbon emission from urban passenger transportation in Beijing

Urban passenger transport significantly contributes to global greenhouse gas emissions, especially in developing countries owing to the rapid motorization, thus making it an important target for carbon reduction. This article established a method to estimate and analyze carbon emission from urban passenger transport including cars, rail transit, taxis and buses. The scope of research was defined based on car registration area, transport types and modes, the stages of rail transit energy consumption. The data availability and gathering were fully illustrated. A city level emission model for the aforementioned four modes of passenger transport was formulated, and parameters including emission factor of electricity and fuel efficiency were tailored according to local situations such as energy structure and field survey. The results reveal that the emission from Beijing’s urban passenger transport in 2012 stood at 15 million tonnes of CO₂, of which 75.5% was from cars, whereas car trip sharing constitutes only 42.5% of the total residential trips. Bus travel, yielding 28.6 g CO₂, is the most efficient mode of transport under the current situations in terms of per passenger kilometer (PKM) emission, whereas car or taxi trips emit more than 5 times that of bus trips. Although a decrease trend appears, Beijing still has potential for further carbon reduction in passenger transport field in contrast to other cities in developed countries. Development of rail transit and further limitation on cars could assist in reducing 4.39 million tonnes CO₂ emission.     

A bi‐level programming approach — Optimal transit fare under line capacity constraints

The fare of a transit line is one of the important decision variables for transit network design. It has been advocated as an efficient means of coordinating the transit passenger flows and of alleviating congestion in the transit network. This paper shows how transit fare can be optimized so as to balance the passenger flow on the transit network and to reduce the overload delays of passengers at transit stops. A bi‐level programming method is developed to optimize the transit fare under line capacity constraints. The upper‐level problem seeks to minimize the total network travel time, while the lower‐level problem is a stochastic user equilibrium transit assignment model with line capacity constraints. A heuristic solution algorithm based on sensitivity analysis is proposed. Numerical example is used to illustrate the application of the proposed model and solution algorithm.     

Stochastic dynamic switching in fixed and flexible transit services as market entry-exit real options

The first analytical stochastic and dynamic model for optimizing transit service switching is proposed for “smart transit” applications and for operating shared autonomous transit fleets. The model assumes a region that requires many-to-one last mile transit service either with fixed-route buses or flexible-route, on-demand buses. The demand density evolves continuously over time as an Ornstein-Uhlenbeck process. The optimal policy is determined by solving the switching problem as a market entry and exit real options model. Analysis using the model on a benchmark computational example illustrates the presence of a hysteresis effect, an indifference band that is sensitive to transportation system state and demand parameters, as well as the presence of switching thresholds that exhibit asymmetric sensitivities to transportation system conditions. The proposed policy is computationally compared in a 24-hour simulation to a “perfect information” set of decisions and a myopic policy that has been dominant in the flexible transit literature, with results that suggest the proposed policy can reduce by up to 72% of the excess cost in the myopic policy. Computational experiments of the “modular vehicle” policy demonstrate the existence of an option premium for having flexibility to switch between two vehicle sizes.     

A linear programming formulation for autonomous intersection control within a dynamic traffic assignment and connected vehicle environment

This paper develops a novel linear programming formulation for autonomous intersection control (LPAIC) accounting for traffic dynamics within a connected vehicle environment. Firstly, a lane based bi-level optimization model is introduced to propagate traffic flows in the network, accounting for dynamic departure time, dynamic route choice, and autonomous intersection control in the context of system optimum network model. Then the bi-level optimization model is transformed to the linear programming formulation by relaxing the nonlinear constraints with a set of linear inequalities. One special feature of the LPAIC formulation is that the entries of the constraint matrix has only {−1, 0, 1} values. Moreover, it is proved that the constraint matrix is totally unimodular, the optimal solution exists and contains only integer values. It is also shown that the traffic flows from different lanes pass through the conflict points of the intersection safely and there are no holding flows in the solution. Three numerical case studies are conducted to demonstrate the properties and effectiveness of the LPAIC formulation to solve autonomous intersection control.     

Short-term forecasting of high-speed rail demand: A hybrid approach combining ensemble empirical mode decomposition and gray support vector machine with real-world applications in China

Short-term forecasting of high-speed rail (HSR) passenger flow provides daily ridership estimates that account for day-to-day demand variations in the near future (e.g., next week, next month). It is one of the most critical tasks in high-speed passenger rail planning, operational decision-making and dynamic operation adjustment. An accurate short-term HSR demand prediction provides a basis for effective rail revenue management. In this paper, a hybrid short-term demand forecasting approach is developed by combining the ensemble empirical mode decomposition (EEMD) and grey support vector machine (GSVM) models. There are three steps in this hybrid forecasting approach: (i) decompose short-term passenger flow data with noises into a number of intrinsic mode functions (IMFs) and a trend term; (ii) predict each IMF using GSVM calibrated by the particle swarm optimization (PSO); (iii) reconstruct the refined IMF components to produce the final predicted daily HSR passenger flow, where the PSO is also applied to achieve the optimal refactoring combination. This innovative hybrid approach is demonstrated with three typical origin–destination pairs along the Wuhan-Guangzhou HSR in China. Mean absolute percentage errors of the EEMD-GSVM predictions using testing sets are 6.7%, 5.1% and 6.5%, respectively, which are much lower than those of two existing forecasting approaches (support vector machine and autoregressive integrated moving average). Application results indicate that the proposed hybrid forecasting approach performs well in terms of prediction accuracy and is especially suitable for short-term HSR passenger flow forecasting.     

Two-phase stochastic program for transit network design under demand uncertainty

This paper develops a reliability-based formulation for rapid transit network design under demand uncertainty. We use the notion of service reliability to confine the stochastic demand into a bounded uncertainty set that the rapid transit network is designed to cover. To evaluate the outcome of the service reliability chosen, flexible services are introduced to carry the demand overflow that exceeds the capacity of the rapid transit network such designed. A two-phase stochastic program is formulated, in which the transit line alignments and frequencies are determined in phase 1 for a specified level of service reliability; whereas in phase 2, flexible services are determined depending on the demand realization to capture the cost of demand overflow. Then the service reliability is optimized to minimize the combined rapid transit network cost obtained in phase 1, and the flexible services cost and passenger cost obtained in phase 2. The transit line alignments and passenger flows are studied under the principles of system optimal (SO) and user equilibrium (UE). We then develop a two-phase solution algorithm that combines the gradient method and neighborhood search and apply it to a series of networks. The results demonstrate the advantages of utilizing the two-phase formulation to determine the service reliability as compared with the traditional robust formulation that pre-specifies a robustness level.     

Implications of climate change for thermal discomfort on underground railways

Hot weather events, ventilation assets, changing passenger demand and service expectations have all caused increased attention on thermal comfort on London’s Tube. This study provides estimates of the future number of days when passengers travelling on sections of the Tube could be subjected to thermal discomfort under future scenarios of climate change, and the potential number of passengers dissatisfied. A risk based methodology is presented, integrating a spatial weather generator modified for urban areas and a thermal comfort model. The study provides an initial assessment of adaptation options by considering the implications of lowering train temperatures by 2 °C and 4 °C to represent saloon cooling. Median results under a 2050 high scenario indicate that all Tube lines assessed could experience near-complete passenger dissatisfaction with the thermal environment in trains in the unlikely event that nothing else were to change. Adaptation aimed at lowering train temperatures has the potential to provide tangible improvements in thermal comfort. However, this was not projected to be sufficient to maintain comfortable thermal conditions for many of the lines in the 2050s under high emission scenarios, requiring a combination of other infrastructure cooling measures to be implemented in parallel.     

Maximizing net benefits for conventional and flexible bus services

Transit ridership is usually sensitive to fares, travel times, waiting times, and access times, among other factors. Therefore, the elasticities of demand with respect to such factors should be considered in modeling bus transit services and must be considered when maximizing net benefits (i.e. “system welfare” = consumer surplus + producer surplus) rather just minimizing costs. In this paper welfare is maximized with elastic demand relations for both conventional (fixed route) and flexible-route services in systems with multiple dissimilar regions and periods. As maximum welfare formulations are usually too complex for exact solutions, they have only been used in a few studies focused on conventional transit services. This limitation is overcome here for both conventional and flexible transit services by using a Real Coded Genetic Algorithm to solve such mixed integer nonlinear welfare maximization problems with constraints on capacities and subsidies. The optimized variables include service type, zone sizes, headways and fares. We also determine the maximum welfare threshold between optimized conventional and flexible services) and explore the effects of subsidies. The proposed planning models should be useful in selecting the service type and optimizing other service characteristics based on local geographic characteristics and financial constraints.     

Activity-based market equilibrium for capacitated multimodal transport systems

Empirical studies have shown that demand for multimodal transport systems is highly correlated with activity schedules of individuals. Nonetheless, existing analytical equilibrium models of multimodal systems have only considered trip-based demand. We propose a new market equilibrium model that is sensitive to traveler activity schedules and system capacities. The model is based on a constrained mixed logit model of activity schedule choice, where each schedule in the choice set is generated with a multimodal extension of the household activity pattern problem. The extension explicitly accounts for both passenger choices of activity participation and multimodal choices like public transit, walking, and vehicle parking. The market equilibrium is achieved with Lagrangian relaxation to determine the optimal dual price of the capacity constraint, and a method of successive averages with column generation finds an efficient choice set of activity schedules to assign flows over the dynamic network load capacities. An example illustrates the model and algorithm, effects similar to Vickrey’s morning commute model can be observed as a special case. A case study of the Oakville Go Transit station access “last mile” problem in the Greater Toronto Area is conducted with 166 survey samples reflecting 3680 individuals. Results suggest that a $10 fixed parking fee at Oakville station would lead to a reduction of access auto share from 54.8% to 49.5%, an increase in access transit share from 20.7% to 25.9%, and a disutility increase of 11% for the of single-activity residents of Oakville.     

On seat capacity in traffic assignment to a transit network

Seating or standing make distinct on‐board states to a transit rider, yielding distinct discomfort costs, with potential influence on the passenger route choice onto the transit network. The paper provides a transit assignment model that captures the seating capacity and its occupancy along any transit route. The main assumptions pertain to: the seat capacity by service route, selfish user behaviour, a seat allocation process with priority rules among the riders, according to their prior state either on‐board or at boarding. To each transit leg from access to egress station is associated a set of ‘service modes’, among which the riders are assigned in a probabilistic way, conditionally on their priority status and the ratio between the available capacity and the flow of them. Thus the leg cost is a random variable, with mean value to be included in the trip disutility. Computationally efficient algorithms are provided for, respectively, loading the leg flows and evaluating the leg costs along a transit line. At the network level, a hyperpath formulation is provided for supply‐demand equilibrium, together with a property of existence and an method of successive averages equilibration algorithm. It is shown that multiple equilibria may arise. Copyright © 2010 John Wiley & Sons, Ltd.     

A schedule-based time-dependent trip assignment model for transit networks

A schedule-based time-dependent trip assignment model for transit networks is presented. First the transit network model is formulated using the schedule-based approach, in which the vehicles are assumed to arrive punctually in accordance with a scheduled time-table. Based on a previously developed time-dependent shortest path algorithm, an all-or-nothing network loading procedure is employed to assign the passenger trips onto the network. Both the passenger demand and scheduled time-table are time-varying. This provides a versatile tool for the evaluation of the performance of transit networks subject to peak period loading. A case study using the Mass Transit Railway System in Hong Kong is given to illustrate the potential applications of the model.     

弹性需求下的轨道交通客流分配模型和算法研究

文章分析了轨道交通客流需求量的影响因素,以拥挤条件下的出行阻抗函数为基础,通过引入弹性需求条件下的轨道交通均衡配流条件,构建了弹性需求的均衡配流模型。根据模型的特点,给出了改进的用于求解弹性需求下的轨道交通均衡配流模型的Frank-wolfe算法。最后通过一个算例说明了算法的有效性和合理性。     

The impact of urban rail transit on commercial property value: New evidence from Wuhan,China

The interaction between rail transit and the urban property market is a vital foundation for planning transit-based policy such as Value Capture and Transit Oriented Development (TOD). Yet only few studies have reported the impact of transit access on commercial property value. This paper presents empirical evidence from Wuhan, China, to enrich the knowledge in the subject area. Spatial autoregressive models were employed to estimate the commercial value capture, based on 676 observations along Wuhan’s metro rail line through the main business districts. Value appreciation was discovered within the 400 m radius of road network distance from Metro stations. The transit access premiums present as two tiers: 16.7% for the 0–100 m core area and approximately 8.0% within the 100–400 m radius. The result demonstrates the potential benefit of adopting value capture and optimising TOD planning to support sustainable urban rail transit investment. Amid rapid urbanisation in China, the evidence reported here could help better inform cities, across the developing world and beyond, of the benefits of adopting rail transit-based policy.     

Commuters’ exposure to PM2.5 and CO2 in metro carriages of Shanghai metro system

Air quality inside transportation carriages has become a public concern. A comprehensive measurement campaign was conducted to examine the commuters’ exposure to PM_2.5 (d_p ⩽ 2.5 μm) and CO₂ in Shanghai metro system under different conditions. The PM_2.5 and CO₂ concentrations inside all the measured metro lines were observed at 84 ± 42 μg/m³ and 1253.1 ± 449.1 ppm, respectively. The factors that determine the in-carriage PM_2.5 and CO₂ concentrations were quantitatively investigated. The metro in-carriage PM_2.5 concentrations were significantly affected by the ventilation systems, out-carriage PM_2.5 concentrations and the passenger numbers. The largest in-carriage PM_2.5 and CO₂ concentrations were observed at 132 μg/m³ and 1855.0 ppm inside the carriages equipped with the oldest ventilation systems. The average PM_2.5 and CO₂ concentrations increased by 24.14% and 9.93% as the metro was driven from underground to overground. The average in-carriage PM_2.5 concentrations increased by 17.19% and CO₂ concentration decreased by 16.97% as the metro was driven from urban to the suburban area. It was found that PM_2.5 concentration is proportional to the on-board passenger number at a ratio of 0.4 μg/m³·passenger. A mass-balance model was developed to estimate the in-carriage PM_2.5 concentration under different driving conditions.     

Real time transit demand prediction capturing station interactions and impact of special events

Demand for public transportation is highly affected by passengers’ experience and the level of service provided. Thus, it is vital for transit agencies to deploy adaptive strategies to respond to changes in demand or supply in a timely manner, and prevent unwanted deterioration in service quality. In this paper, a real time prediction methodology, based on univariate and multivariate state-space models, is developed to predict the short-term passenger arrivals at transit stations. A univariate state-space model is developed at the station level. Through a hierarchical clustering algorithm with correlation distance, stations with similar demand patterns are identified. A dynamic factor model is proposed for each cluster, capturing station interdependencies through a set of common factors. Both approaches can model the effect of exogenous events (such as football games). Ensemble predictions are then obtained by combining the outputs from the two models, based on their respective accuracy. We evaluate these models using data from the 32 stations on the Central line of the London Underground (LU), operated by Transport for London (TfL). The results indicate that the proposed methodology performs well in predicting short-term station arrivals for the set of test days. For most stations, ensemble prediction has the lowest mean error, as well as the smallest range of error, and exhibits more robust performance across the test days.     

Statistical modeling of Electric Vehicle electricity consumption in the Victorian EV Trial,Australia

The market share of Electric Vehicles (EVs), an attractive alternative to conventional vehicles, is expected to exceed 30% of all vehicles by 2033 in Australia. Although the expected EV uptake may place greater burdens on electricity networks, the potential impacts contributed by different EV user categories and vehicle models to peak loads at different times during the day are not well understood. This paper addresses the issue through statistical analysis of the charge events in the Victorian EV Trial in Australia as well as modeling the charging behaviors according to participant categories and vehicle models. The analysis was performed on 4933 charge events that were recorded by both private and public Electric Vehicle Supply Equipment. In total, these events consumed over 33 MW h of energy over 12,170 h by the 178 trial participants, out of which about 70% were household participants while the others were fleet participants. Based on a range of EV uptake scenarios and modeled charging behaviors from the trial, the power demand in the summer of 2032/33 was estimated for all of Victoria. The results of the simulations show that the broad scale uptake of EVs produces a relatively small increase in overall power demand (estimated to be between 5.72% and 9.79% in 2032/33).     

Estimation of origin‐destination matrices for a multimodal public transit network

This paper presents a procedure for the estimation of origin‐destination (O‐D) matrices for a multimodal public transit network. The system consists of a number of favored public transit modes that are obtained from a modal split process in a traditional four‐step transportation model. The demand of each favored mode is assigned to the multimodal network, which is comprised of a set of connected links of different public transit modes. An entropy maximization procedure is proposed to simultaneously estimate the O‐D demand matrices of all favored modes, which are consistent with target data sets such as the boarding counts and line segment flows that are observed directly in the network. A case study of the Hong Kong multimodal transit network is used to demonstrate the effectiveness of the proposed methodology.