Evacuation travel behavior in regret minimization or utility maximization rules? Evidence from emergency context

Abstract

The aim of this paper is to compare Random Regret Minimization (RRM) and Random Utility Maximization (RUM) paradigms in emergency evacuation decision application. Three models including utility model, regret model, and hybrid utility-regret model are formulated, respectively. A Stated Preference (SP) survey on evacuation route choice under emergency was conducted to obtain the evacuees' trade-off among average travel time, travel time uncertainty, perceived road damage probability, and perceived service level. The estimation reveals that regret-based model performs better than utility-based model. The differences between RRM and RUM in emergency context are substantial in model fit when compared with previous empirical results under non-emergency state. Specifically, the model also explains the influence of determining factors related to evacuees' socio-demographics on their route decisions. Further, considering the road damage probability as a predominant factor, Value of Damage (VOD) is proposed to reveal the unobserved regret aversion psychology. Direct elasticity and cross elasticity are calculated for the hybrid utility-regret model. Regret increases in the emergency scenario, supporting that the advantage of RRM paradigm to explain evacuation route choices. The study is expected to give a better understanding of evacuation route choice behavior for emergency actions.