Agent-Based Modeling and Evacuation Planning

Abstract

In evacuation planning, it is advantageous for community leaders to have a thorough understanding of the human and geophysical characteristics of a community, be able to anticipate possible outcomes of different response and evacuation strategies under different situations, inform the general public, and develop a set of evacuation plans accordingly. In order to achieve this goal, evacuation managers in a community can use computer modeling techniques to simulate different ‘what-if’ scenarios, use the results from these simulations to inform the public, and generate different evacuation plans under different circumstances. The complexity associated with evacuation planning in an urban environment requires a computer modeling framework that can incorporate a number of factors into the modeling process. These factors include the nature of the disaster in question, the anticipated human behavioral patterns in the evacuation process, the unique geography and transportation infrastructure in a given area, the population distribution in the area, the population dynamics over different time periods, and the special needs of different population groups, to name a few. Agent-Based Modeling (ABM) provides a general approach that can be used to account for these factors in the modeling and simulation process. In this chapter, the authors provide an overview of agent-based modeling and simulation, illustrate how agent-based modeling and simulation were used in estimating the evacuation time for the Florida Keys, and report some preliminary results in planning a hypothetical route for evacuating the elderly from a nursing home on Galveston Island, Texas, based on network dynamics during an evacuation.