Abstracts

Reliability is an important aspect of any transportation system, and there has been a substantial amount of research to bring it to the point where it can be accurately incorporated into established forecasting and appraisal frameworks. At the same time, emerging data sources, such as public transport smartcards, provide opportunities to understand more about the reliability of a particular transport system. This thesis conducts research on reliability using smartcard data. In the first instance the thesis provides a critique of the Mean-Variance framework for the treatment of transport reliability and finds room for adaptation. The thesis also provides a review of empirical research that estimates Mean-Variance variables and parameters, and finds evidence of methodological issues.In response to these issues, the thesis utilises smartcard data to investigate Mean-Variance in three ways. The key element is the development of an alternative method of estimating a value of reliability, treating smartcard data as a Revealed Preference data source and combining it with established discrete choice methods. The second element uses the smartcard data to identify the factors affecting reliability levels through estimation of a linear regression model. The third strand of investigation employs the data to understand more about possible alternative measures of reliability and compares the underlying utility function of Mean-Variance with a second reliability framework. The thesis therefore demonstrates that the application of public transport smartcard data has the potential to yield insights in the field of transport reliability. In particular, it establishes how this data source might be used to estimate the value of reliability. With development, it may have the potential to forecast future reliability levels. Application of the data also supports the status quo of utilising the standard deviation as an indicator of reliability.