Summary of the Master Thesis
“The Costs of Bus Unreliability”
for M.Sc. transport and logistic
This work deals with the estimation of the costs that passengers experience from travel time unreliability, and extends the methodology to evaluate benefits passengers would experience from improvements in travel time reliability. Absence of such estimate means that it is problematic to include reliability as a factor in an assessment of any public transport improvement proposal. Consequently, the reliability improvement may be delayed or disregarded due to the lack of justification for the costs.
The challenge in measuring and obtaining the value of reliability is largely due to the discrepancies between perception of reliability by passengers and planners. Passenger experience of unreliability includes several ways the reliability translates to their travel components: waiting time, variability of travel time, as well as earliness and lateness at the trip destination. They value improvements in reliability as increased punctuality at work or school. However, such indicators are difficult to use as performance guidance for operators due to the inability to measure and monitor them. Typically, the operators would measure and present reliability by statistics of headway distribution, such as mean headway and variance of headway.
This work aims to translate the change of operational indicators of reliability to change in travel components for passengers like their experienced waiting time, earliness and lateness. To enable this transition, headway data of Singapore buses had to be studied for the goodness-of-fit of statistical distributions. This was successfully accomplished, and best fitting distribution implemented in the translation framework. Having this transition model in place, it is possible to estimate the passenger benefit from improvements of operational indicators, if passenger valuation of travel components is known. The output of this model is condensed in two indicators: Value of Service Headway and Value of Service Reliability, which stand for the passenger benefit from improvements in mean headway and standard deviation of headway respectively. If desired, the indicators can be expressed in monetary terms.
Thus, this work reduces a major obstacle in quantifying and including reliability improvements in project pre-assessment. The framework built in this work can be used to estimate passenger benefits from bus service improvement proposals such as introducing more buses and improving the stability and reducing variability of headways.
Scope: This work addresses the translation process of passenger valuations to operationally useful indicators. It does not include implementation of the methodology, because that would require obtaining passenger valuations of travel components. This goal will be pursued in future work. The adjustment of the framework is done for Singapore bus services, and bus headway distribution is tested in morning peak hours.
Objective: Research objective is development of framework to translate passenger valuations of unreliability into operational indicators. This objective can be separated in two sub-objectives – the choice of utility function for the valuation of reliability from the passenger’s perspective, and the interpretation of the components of the utility function in terms of indicators commonly used in practice.
Approach: The first sub-objective could be addressed by literature research and comparison of existing utility functions for reliability measurement. The second objective required a practical part of distribution testing. Smart card data were used to obtain data about bus headways and statistical methodology was used to test the fit of the data to several distributions, which were selected based on recommendations about suitable distributions for headways from literature. Implementation of the best distribution, as well as sensitivity analysis of results was performed.