Abstract An efficient evacuation management plan plays a critical role in reducing loss of life and property during disasters. Developing such plans requires a clear understanding of evacuees’ heterogeneous behavioural preferences, yet existing studies do not sufficiently address this issue. This paper presents a scenario-based Stated Preference (SP) survey with fixed choice tasks to examine route choice behaviour under both normal travel conditions and flood evacuation contexts. The survey collects data from New South Wales, Australia and classifies respondents into nine preference categories based on risk attitudes. Based on the SP data, this study develops and estimates three behaviourally motivated logit route choice models, including a perceived utility-based model and two regret based models grounded in Random Regret Minimisation (RRM). The results show that among the nine categories of preferences, the risk-seeking in normal and disaster group had the lowest values for the three parameter estimates, while the risk-averse in normal and disaster group had the highest value for the three parameter estimates. This study makes three specific contributions to the literature on evacuation route choice behaviour. First, it provides empirical evidence on how regret aversion and risk preferences jointly shape evacuation route choice decisions, revealing that risk-seeking groups exhibit systematically lower perceived utility scale values, regret aversion, and risk aversion parameters compared to risk-averse groups. Second, it develops and compares three behaviourally motivated route choice models, demonstrating that incorporating regret- and risk-related behavioural mechanisms provides a richer characterisation of evacuation route choice than utility-based models alone. Third, it offers empirically grounded policy insights for flood evacuation management, suggesting that uniform evacuation guidance may be insufficient and that targeted communication strategies tailored to different risk preference profiles may improve evacuation compliance and efficiency.
Xu et al. (Sat,) studied this question.