Service reliability and urban public transport design

The last few decades have shown a substantial increase in personal mobility. Urban traffic and transport volumes have been increasing for years. However, the share of public transport in this mobility growth did not change much and still remains rather limited. To ensure the accessibility and liveability of our cities for future generations, however, a substantial quality leap in public transport is necessary. This will facilitate a desired modal shift from car traffic towards public transport, which is safer, cleaner and produces less congestion. In this thesis, we demonstrate that several promising opportunities exist to improve service reliability (i.e. the certainty of service aspects compared to the schedule as perceived by the user), being one of the most important quality aspects of public transport. Literature shows that in urban public transport substantial attention is given to ways to improve service reliability at the operational level. It is not clear how and to what extent strategic and tactical design decisions in public transport systems might affect service reliability. Only traffic light priority and exclusive lanes are considered during the planning of urban public transport in order to improve the level of service reliability. We expect that more instruments at these planning levels enable high-quality services at the operational level, especially with regard to service reliability. In this thesis, we present several planning instruments that facilitate enhanced service reliability. In addition, we show forecasting tools we developed and we introduce a new indicator that expresses the impacts of service reliability more effectively than traditional indicators, namely the additional travel time per passenger. This way, the assessment of public transport benefits will be substantially improved, thereby enabling cost-effective quality improvements. We show the impacts of unreliable services on passengers, being average travel time extension, increased travel time variability and a lower probability of finding a seat in the vehicle. We demonstrate how actual vehicle trip time variability (i.e. service variability) affects service reliability and passenger travel time. In order to gain insights into the mechanisms between these two aspects, we performed research based on empirical data of the public transport system in The Hague. Furthermore, we conducted an international survey of service reliability. Several traditional quantifications of service reliability are presented, such as punctuality and regularity. We demonstrate the shortcomings of these traditional indicators, namely a lack of attention for passenger impacts. Traditional indicators focus too much on the supply side of public transport, which does not allow a proper analysis of passenger effects. To deal with the shortcomings of traditional indicators, we developed a new indicator, being the average additional travel time per passenger. This indicator translates the supply-side indicators, for instance punctuality, into the additional travel time that a passenger on average needs to travel from the origin to the destination stop due to service variability. The average additional travel time may be calculated per stop or per line and enables explicit consideration of service reliability in cost-benefit calculations, since the level of service reliability may be translated into regular travel time. In our research we demonstrate that our indicator enables optimization of network and timetable planning and that the use of traditional indicators may lead to conflicting conclusions in terms of service reliability. We also demonstrate the benefits of using reliability buffer time (RBT, as described by Furth and Muller 2006) as an indication of the effects of uncertain arrivals for passengers (i.e. the variability of the average additional travel time per passenger). To improve service reliability through enhanced network and timetable design, we selected five planning instruments by analysing the causes of service variability. The main external causes are the weather, other traffic, irregular loads and passengers’ behaviour. Other public transport, driver behaviour, schedule quality and network and vehicle design are the main internal causes of unreliability. Since the arrival pattern of passengers is very important when calculating service reliability effects, we performed a passenger survey in The Hague. It showed that passengers tend to arrive at random at their departure stop if scheduled headways are 10 minutes or less. In the case of longer headways, passengers on average plan their arrival about 2 minutes prior to the scheduled vehicle departure time. Applying instruments during the planning stages will reduce the impacts of these causes. At the strategic level, these instruments are: - Terminal design; The configuration and number of tracks and switches at the terminal determines the expected vehicle delay and thus service reliability. - Line length; The length of a line is often related to the level of service variability and thus service reliability. - Line coordination. Multiple lines on a shared track may offer a higher level of service reliability than one line (assuming equal frequencies). The following instruments may be applied at the tactical level: - Trip time determination; In long-headway services, scheduled vehicle departure times at the stop, derived from scheduled trip times, determine the arrival pattern of passengers at their departure stop. Adjusting the scheduled trip time may affect the level of service reliability and passenger waiting time. - Vehicle holding. Holding early vehicles reduces driving ahead of schedule and increases the level of service reliability. The design of the schedule affects the effectiveness of this instrument. The terminal design instrument relates to (new) rail lines with tail tracks as terminal or short-turning facilities. For high-frequency, distributed lines, we recommend compact tail tracks with double crossovers directly after the stop. Concerning (new) lines with a clear break point in passenger pattern, we recommend to split the line or to apply holding points. For long-headway services we propose to use the 35-percentile value for scheduled trip time. And if parts of lines are very crowded, we suggest investigating the effects of coordination. A tentative cost-effectiveness assessment showed that the tactical instruments (trip determination and vehicle holding) are cost-effective in almost every case. Their benefits are substantial and the costs are nil. These instruments should be considered in the design of every public transport system. However, the vehicle holding instrument is only beneficial if the passenger pattern has a clear break point and trip time determination only is relevant in long-headway services. It is presented that strategic instruments have considerable benefits as well. Optimized terminal design enables enhanced service reliability. Coordination and shorter lines may result in reduced passenger travel times as well. However, these instruments may look costly due to necessary additional infrastructure and or (occasionally) additional vehicles. We demonstrated that the costs are limited in relation to the potential welfare benefits. We roughly estimated the costs of unreliability at \u80 12 million per year in The Hague and we estimated the potential savings at \u80 8 million per year by applying the five planning instruments we analysed in our research. The estimated costs of these instruments are assessed to be only a part of the benefits with a maximum of \u80 3 million per year, showing the added value of the instruments. The results of our international survey show that similar results are achievable in other cities as well. In this thesis we presented planning instruments that facilitate enhanced service reliability. To achieve a higher level of service reliability in practice, we recommend considering service reliability explicitly in the design of infrastructure networks, service networks and timetables using our developed control framework and tools. Service reliability effects should be incorporated in a sophisticated way into cost-benefit analyses of public transport projects, using the average additional travel time per passenger we introduced. This way, welfare gains and additional revenues may be calculated. Optimized strategic and tactical design improves service reliability and also simplifies the operational process with regard to service reliability. Enhanced service planning will allow passengers to benefit from improved service reliability tomorrow!Transport and PlanningCivil Engineering and Geoscience