Can Agent-Based Models Simulate Travel Behaviour During Disruptive Events?

As transport infrastructure ages, it is affected by extreme weather conditions or is overused, its resilience decreases and its risk of damage increases. But what happens to the users of that infrastructure once it is no longer in service? Such events could impact travel demand and, as a cascading effect, the economy, society, and environment. This paper presents the results of a MATSim Agent-based Modelling (AgBM) scenario, analysing the impact of the 4-year Tyne Bridge restoration work in the North East of England (UK). The consequences of reducing the maximum speed and traffic flow in the affected area by 50% are simulated to identify the agents’ behavioural changes. Results reflect small spatio-temporal differences in daily vehicle counts at specific locations (up to 7% in most areas) compared to ground truth data, and minimal modal shift (2%). Furthermore, alternative routes taken by agents, increases in average travel times and distances for specific population groups, the effects of ‘disappearing traffic’ and transport network resilience are analysed. This work demonstrates the value of AgBMs for simulating the effects of transport disruptions, as they can provide highly detailed representation of multimodal human behaviours at the individual scale, unlike other traditional transport models.