The Importance of Modelling Realistic Human Behaviour When Planning Evacuation Schedules

In planning evacuation schedules various projects work on devising optimal schedules for evacuation of the population. These assume that people do as they are asked in terms of when and by what route they leave the area. However, we know from numerous case studies, that this is not so. People have their own priorities and concerns and will address these before doing as they are asked. We know from many emergency situation examples that a key behaviour that people engage in is to assemble the family group, and to check on relatives or close friends. These behaviours are highly likely to affect how an optimal schedule that is provided plays out in practice. In this work we obtain an actual optimised evacuation schedule for a large area of over 35,000 households, which is prone to wildfires. We then compare an agent based simulation that adheres to this schedule, with ones that use it as a base, but where some of the agents first engage in the behaviours mentioned. We analyse the differences on 17 different configurations, exploring both the statistical significance of differences, as well as the importance of the differences for ensuring successful evacuation.