From Social Networks to Evacuation Traveler Decision Making: Exploratory Departure Time Choice Modeling and Simulation

Natural and man-made disasters lead to massive human, economic, and social losses. The frequency and severity of these disasters creates the need for more effective evacuation strategies. Recent events, namely Hurricane Katrina (the single most expensive and among the five deadliest natural disasters in U.S. history), have demonstrated that human behavior during an evacuation is complex and difficult to predict accurately. Analysis of human behavior during similar extreme events has also shown that social interactions have a significant impact on decision making logic. As such, the objective of this paper is to account for social interactions and the role of an individuals' social network in the transportation evacuation model. The authors begin on the microscopic level by developing a method for forming links between individuals based on common characteristics. Next, they move to the macroscopic level and utilize a latent space approach to determine the influence individuals have on one another within a social network. Finally, the microscopic and macroscopic data are combined within the framework of a choice model in order to transform the social data into a binary decision variable; either to evacuate or not to evacuate. These decision variables form a demand loading scheme which can be used to simulate evacuation scenarios in a transportation simulation software.