Activity-based Modeling and Microsimulation of Emergency Evacuations

Most existing emergency evacuation studies assume evacuees to evacuate from their residence locations. However, depending on the time of day, people's movements are constrained not only by their workplaces, but also the necessity of picking up family members. Family member interactions, a typical activity during an evacuation, can greatly affect the evacuation process. Activity-based modeling has been applied to estimate daily traffic demand widely. However, only limited research has been reported to incorporate the activity component in examining evacuation processes, particularly for mass evacuations at a micro-scale. Under this context, this study aims to analyze the activity-based travel pattern and its impact on emergency evacuations in the case of a hypothetical emergency evacuation of Galveston Island, Texas. In the study, one typical type of daily trip, picking up school-age children, is considered. All households with school-age children are assumed to pick up their children first in an evacuation. Trip chains are defined to represent the movements from workplaces to schools and then to destinations. This study employs agent-based microsimulation techniques to model the evacuation process at the individual-driver level. The simulation results suggest that the overall evacuation time may not be significantly affected when the trips of picking-up school-age children are considered in the case of Galveston Island. However, the average travel and delay time of individual vehicles may increase dramatically, which suggests the occurrence of considerable congestions during the evacuation. The findings demonstrated the importance of considering activity-based trips in evacuations.