Magnus Frestad Nygaard
Department of Civil and Transport Engineering Norwegian University of Science and Technology, Norway
Trude Tørset
Department of Civil and Transport Engineering Norwegian University of Science and Technology, Norway
1 Need for better modelling of waiting time
In todays transport models the waiting time for public transport travels is set to half the time between departures. This is a good representation, or at least a good-enough estimation, of real waiting time for small headways. However, when the headway is getting larger, this estimation is starting to differ from the real waiting time which public transport passengers would experience.
When the headways are large, the passengers tend to plan their arrival to the platform based on the departure time according to the trip table. This results in that the platform waiting time is less than half the time between departures. It is true that the travelers
experience a hidden waiting time, because they cannot choose to travel whenever they want, in addition to the platform waiting time. However, this part of the waiting time should not have the same travel costs as the platform waiting time. Which it has in todays transport models. In addition, delays or capacity issues in the public transport service might influence the arrival strategy and the actual waiting time, and these matters are not dealt with in transport models.
This results in that the calculated change in demand, due to changes in the public transport service that will affect waiting time, will not be correct. The erroneous estimate of demand changes might be large in cities where the public transport service operate with low departure frequency on all, or most of, the lines, and this might lead to a supply of public transport services that does not match the true demand. A situation like this will occur in all Norwegian cities, and other small and medium sized cities worldwide.
406
2 Objectives
Comparing reported waiting time with waiting time used by the transport models has shown that the models are misrepresenting the waiting time. The common model approach is
assumptions of uniform arrival distribution that is not supported by empirical studies, least not in a low frequent public transport system, which is the situation in most Norwegian cities.
Our approach will study real life travel behavior, in order to reveal public transport travelers waiting time strategy. The main goal is to reveal how the strategy vary dependent on the public transport service. More specifically, how travelers value the open and the hidden waiting time, the limit of low frequency before the passengers start using the trip table, and the difference in waiting time values when the waiting is regular, compared to when the waiting is caused by delays.
3 Methodology
There will be done registrations for when travelers arrive at bus stops, in order to see how they plan their trips according to scheduled bus departure times. The registrations are
conducted manually on site, with use of a self-designed computing program that can store the exact time of arrivals of both passengers and vehicles. The registrations will take place mostly during the morning peak-hours, on several bus stops in residential areas, which is served by a single bus line.
The arrival pattern will reveal some basic information of passengers waiting time strategy, but in order to find information that is more specific, one or more surveys will be conducted. The survey(s) goal is to find how travelers value their open, hidden, and excess waiting time, and if there is some logical connections between waiting strategy and age, sex, and travel purpose.
4.1 Expected results
Based on the results from this study it should be possible to see how waiting time strategies vary among travelers with different access to public transport. This will show how waiting time, both open and hidden, should be implemented in the transport models.
The study’s aim to clarify that public transport services with longer headways should not make use of today’s implementations of waiting time (W=h/2) in transport models, in order to give accurate estimations for public transport demand.
4.2 Preliminary results
Registrations conducted on different bus stops that is served by a single bus line with 10, 15, and 20 minutes frequency, have clearly shown that passengers arrivals to bus stops are not random and uniform. The results shows that the average waiting time for public transport passengers are considerably smaller than half the time between departures, and that the arrival pattern does vary between different bus services.