Train dwell time models – development in the past forty years

The accuracy of dwell time estimation is crucial for both tactical and operational practices of public transport. This paper aims to make a systematic review of the dwell time models that have evolved over the past 40 years. The scope of this study is limited to the dwell time models pertaining to the passenger rail. Studying the literature concerning train dwell time models, similarities and differences were analysed and discussed based on the modelling approaches and assumptions made in the development of the models. For instance, models were categorised based on modelling philosophy, time-period, key variables involved, data collection and validation methods. Through the comparison, common interests and future trends were then identified. It is found that there is no perfect model that fits all scenarios. The best outcome relies on the effort of choosing the most appropriate model, calibrating the parameters, making some ad-hoc adjustment and continual improvements.