Abstract
Existing day-to-day models assume that travellers make decisions based on traffic conditions from previous days rather than those on the actual travel day. However, this approach cannot capture frequent within-day updates of travel information, significantly influencing both day-to-day and within-day dynamics. To address this, we proposed the concept of a within-day decision-making framework and developed a mathematically simple and concise Markovian-based model. Additionally, to characterise and quantify the impacts of the proposed model, we introduced three multi-faceted measures: the convergence speed towards the stationary distribution, its key characteristics, and the within-day adaptation capability, all designed to reflect the stability and resilience of the transport system. We conducted numerical experiments on the departure time choice problem in both a single-link network and the Sioux Falls network. The results clearly demonstrate the impacts of the proposed decision-making framework through the proposed measures, particularly highlighting its effects on the stability and resilience of the transport system.