Data driven public transportation delay modelling

Transportation Research Record: Journal of the Transportation Research Board, 2018

Train operations are subject to stochastic variations, reducing service punctuality and thus the quality of service (QoS). Models of such variations are needed to evaluate and predict the potential impact of disturbances and to avoid service punctuality reduction in train service management and timetabling. In this paper, through a case study of the Wuhan–Guangzhou (WH–GZ) high-speed rail (HSR), we show how a wealth of train operation records can be used to model the stochastic nature of train operations at each level, section and station. Specifically, we examine different distribution models for running times of individual sections and show that the Log-logistic probability density function is the best distributional form to approximate the empirical distribution of running times on the specified line. Next, we show that the distribution of running times in each section can be used to accurately infer arrival delays. Consequently, we construct the underlying analytical model and d...

2014

Bus dwell time is traditionally considered as a function of the number of alighting and boarding passengers plus an amount of time for door opening and closing. This approach describes the delays caused to a bus at any bus stop; however, these could be less accurate for busway stations where more than one loading area is available. This is because these traditional dwell time models do not account for the crowd phenomenon of the busway station platform. This paper discusses the differences in boarding process at a bus stop and at a busway station. And based of these differences, a new methodology for estimation of bus dwell time at busway stations is proposed. For this study, data collection was performed at a busway station on the South East Busway in Brisbane, Australia and at suburban bus stops. Firstly, the sequence of activities performed by a passenger from viewing the desired bus to boarding the bus was divided into two parts: passenger -bus interface and passenger bus interaction. Later the effect of platform crowding on passenger -bus interface and passenger bus interaction was studied. Finally a dwell time model for buses at busway stations was developed. Based on the analysis, this paper concluded that at a busway station where boarding is predominant, an increase in platform crowd increases the passenger -bus interface duration. This increase in passenger -bus interface duration then leads to lost time for buses and hence increases the bus dwell time.

Procedia - Social and Behavioral Sciences, 2012

Along a transit line, vehicle traffic and passenger traffic are jointly subject to variability in travel time and vehicle load hence crowding. The paper provides a stochastic model of passenger physical time and generalized time, including waiting on platform and in-vehicle run time from access to egress station. Five sources of variability are addressed: (i) vehicle headway which can vary between the stations provided that each service run maintains its rank throughout the local distributions of headways; (ii) vehicle order in the schedule of operations; (iii) vehicle capacity; (iv) passenger arrival time; (v) passenger sensitivity to quality of service. The perspective of the operator, which pertains to vehicle runs, is distinguished from the user's one at the disaggregate level of the individual trip, as in survival theory. Analytical properties are established that link the distributions of vehicle headways, vehicle run times, passenger wait times, passenger travel times, and their counterparts in generalized time, in terms of distribution functions, mean, variance and covariance. Many of them stem from Gaussian and log-normal approximations.

Public Transport, 2020

Among the many ways to improve a transit system is a reduction in travel time as experienced by the passenger. Hence, passenger waiting times remain a topic of interest among transit planners. In this study, the effects of transit vehicle delays on passenger waiting time is investigated, as well as the effects of transfer status, boarding location, time of day, and rider travel frequency. The data used in this study were collected using automatic fare collection (AFC) and automatic vehicle location (AVL) technology. A trip chaining algorithm is used to infer the trajectory of each passenger, and as a result produce measures of passenger waiting time and vehicle delay. An analysis of an arterial Bus Rapid Transit (aBRT) line in Saint Paul, Minnesota reveals a waiting time model consistent with previous literature, a positive relationship between vehicle delay and passenger waiting time, and an insignificant relationship between transfer status and passenger waiting time. Finally, a simple model relating waiting time and vehicle delay is provided for the purpose of transit planning and waiting time estimation.

Transportation Research Record: Journal of the Transportation Research Board, 2012

This paper proposes a bus dwell-time model obtained by means of a robust statistical evaluation of boarding passenger data at stops. This model is a change from the generally accepted linear model (i.e., dwell time increases in a fixed rate of time per passenger). The statistical analysis proves the validity of the potential model. This model was derived from a large number of observations on Line 27 of the Transports Municipal Company in Madrid, Spain, and validated with observations from another line (Line 70). The data were gathered by observers who could attest to the influence of occasional incidents in the boarding process. This model can be used to evaluate line capacity more accurately. The analyzed lines were main urban routes with high passenger demand requiring the use of articulated buses. These lines were selected according to two basic criteria: routes with high demand that guarantee good results and routes with similar vehicle types with an onboard payment method. A l...

Transportation Research Record, 2007

Transportation Science, 2013

We consider curbside bus stops of the kind that serve multiple bus routes, and that are isolated from the effects of traffic signals and other stops. A Markov chain embedded in the bus queueing process is used to develop steady-state queueing models for two special cases of this stop type. The models estimate the maximum rate that buses can arrive to, and be served by the stop, and still satisfy a specified target of average bus delay. These models can be used to determine, for example, a stop's suitable number of bus berths, given the bus demand and the specified delay target. The solutions for the two special cases are used to derive a closed-form, parsimonious approximation model for general cases. Our approximations closely match simulations for a range of conditions that arise in real settings. And the approximations unveil how suitable choices for the number of bus berths are influenced by both, the variation in the time that buses spend serving passengers at the stop, and the specified delay target. The models further show why the proxy measure that is used for the delay target in other bus-stop studies is a poor one.

2019

The paper analyses dwell time delays for commuter trains in Stockholm and Tokyo. In both cities, small dwell time delays of at most five minutes make up around 90% of the total delays. Therefore, it is valuable to understand and deal with these disturbances. To this end, we use high resolution data on dwell times and passenger counts from both countries over the last several years. We find that these data alone can explain about 40% of the variation in dwell time delays and produce simple models which can be used in practice to assign more appropriate dwell times. A change of 15 passengers per car, in Tokyo translates to a delay of about one second. For every 10 remaining passengers per door in Stockholm, the delay increases by about one second, and one boarding or alighting passenger per door corresponds to about 0.4 seconds of delay. We also find that trains in Tokyo are much more congested than in Sweden, and that at most stations in the latter, the exchange of passengers is modest. In both cities, the range of dwell time delays is quite narrow, with between 40 and 50 seconds separating the 5 th and 95 th percentiles. This indicates further that most delays, by far, are very small, and that even small adjustments to dwell times can make a big difference in the overall picture. To facilitate such improvements, key stakeholders and practitioners are closely involved with the research.

2018 21st International Conference on Intelligent Transportation Systems (ITSC), 2018

India runs the fourth largest railway transport network size carrying over 8 billion passengers per year. However, the travel experience of passengers is frequently marked by delays, i.e., late arrival of trains at stations, causing inconvenience. In a first, we study the systemic delays in train arrivals using norder Markov frameworks and experiment with two regressionbased models. Using train running-status data collected for two years, we report on an efficient algorithm for estimating delays at railway stations with near accurate results. This work can help railways to manage their resources, while also helping passengers and businesses served by them to efficiently plan their activities.