Hurricane Evacuation Modeling Using Behavior Models and Scenario-Driven Agent-based Simulations

Yuan Zhu, Kun Xie, Kaan Ozbay, Hong Yang

Research output: Contribution to journalConference article

Abstract

Transportation modeling and simulation play an important role in the planning and management of emergency evacuation. It is often indispensable for the preparedness and timely response to extreme events occurring in highly populated areas. Reliable and robust agent-based evacuation models are of great importance to support evacuation decision making. Nevertheless, these models rely on numerous hypothetical causal relationships between the evacuation behavior and a variety of factors including socio-economic characteristics and storm intensity. Understanding the impacts of these factors on evacuation behaviors (e.g., destination and route choices) is crucial in preparing optimal evacuation plans. This paper aims to contribute to the literature by integrating well-calibrated behavior models with an agent-based evacuation simulation model in the context of hurricane evacuation. Specifically, discrete choice models were developed to estimate the evacuation behaviors based on large-scale survey data in Northern New Jersey. Monte-Carlo Markov Chain (MCMC) sampling method was used to estimate evacuation propensity and destination choices for the whole population. Finally, evacuation of over a million residents in the study area was simulated using agent-based simulation built in MATSim. The agent-based modeling framework proposed in this paper provides an integrated methodology for evacuation simulation with specific consideration of agents' behaviors. The simulation results need to be further validated and verified using real-world evacuation data.

Original languageEnglish (US)
Pages (from-to)836-843
Number of pages8
JournalProcedia Computer Science
Volume130
DOIs
StatePublished - Jan 1 2018
Event9th International Conference on Ambient Systems, Networks and Technologies, ANT 2018 - Porto, Indonesia
Duration: May 8 2018May 11 2018

Fingerprint

Hurricanes
Markov processes
Decision making
Sampling
Planning
Economics

Keywords

  • agent-based simulation
  • demand modeling
  • Hurricane evacuation
  • MATSim

ASJC Scopus subject areas

  • Computer Science(all)

Cite this

Hurricane Evacuation Modeling Using Behavior Models and Scenario-Driven Agent-based Simulations. / Zhu, Yuan; Xie, Kun; Ozbay, Kaan; Yang, Hong.

In: Procedia Computer Science, Vol. 130, 01.01.2018, p. 836-843.

Research output: Contribution to journalConference article

Zhu, Yuan ; Xie, Kun ; Ozbay, Kaan ; Yang, Hong. / Hurricane Evacuation Modeling Using Behavior Models and Scenario-Driven Agent-based Simulations. In: Procedia Computer Science. 2018 ; Vol. 130. pp. 836-843.
@article{79d8577d30ed46fa92febcc466e29c96,
title = "Hurricane Evacuation Modeling Using Behavior Models and Scenario-Driven Agent-based Simulations",
abstract = "Transportation modeling and simulation play an important role in the planning and management of emergency evacuation. It is often indispensable for the preparedness and timely response to extreme events occurring in highly populated areas. Reliable and robust agent-based evacuation models are of great importance to support evacuation decision making. Nevertheless, these models rely on numerous hypothetical causal relationships between the evacuation behavior and a variety of factors including socio-economic characteristics and storm intensity. Understanding the impacts of these factors on evacuation behaviors (e.g., destination and route choices) is crucial in preparing optimal evacuation plans. This paper aims to contribute to the literature by integrating well-calibrated behavior models with an agent-based evacuation simulation model in the context of hurricane evacuation. Specifically, discrete choice models were developed to estimate the evacuation behaviors based on large-scale survey data in Northern New Jersey. Monte-Carlo Markov Chain (MCMC) sampling method was used to estimate evacuation propensity and destination choices for the whole population. Finally, evacuation of over a million residents in the study area was simulated using agent-based simulation built in MATSim. The agent-based modeling framework proposed in this paper provides an integrated methodology for evacuation simulation with specific consideration of agents' behaviors. The simulation results need to be further validated and verified using real-world evacuation data.",
keywords = "agent-based simulation, demand modeling, Hurricane evacuation, MATSim",
author = "Yuan Zhu and Kun Xie and Kaan Ozbay and Hong Yang",
year = "2018",
month = "1",
day = "1",
doi = "10.1016/j.procs.2018.04.074",
language = "English (US)",
volume = "130",
pages = "836--843",
journal = "Procedia Computer Science",
issn = "1877-0509",
publisher = "Elsevier BV",

}

TY - JOUR

T1 - Hurricane Evacuation Modeling Using Behavior Models and Scenario-Driven Agent-based Simulations

AU - Zhu, Yuan

AU - Xie, Kun

AU - Ozbay, Kaan

AU - Yang, Hong

PY - 2018/1/1

Y1 - 2018/1/1

N2 - Transportation modeling and simulation play an important role in the planning and management of emergency evacuation. It is often indispensable for the preparedness and timely response to extreme events occurring in highly populated areas. Reliable and robust agent-based evacuation models are of great importance to support evacuation decision making. Nevertheless, these models rely on numerous hypothetical causal relationships between the evacuation behavior and a variety of factors including socio-economic characteristics and storm intensity. Understanding the impacts of these factors on evacuation behaviors (e.g., destination and route choices) is crucial in preparing optimal evacuation plans. This paper aims to contribute to the literature by integrating well-calibrated behavior models with an agent-based evacuation simulation model in the context of hurricane evacuation. Specifically, discrete choice models were developed to estimate the evacuation behaviors based on large-scale survey data in Northern New Jersey. Monte-Carlo Markov Chain (MCMC) sampling method was used to estimate evacuation propensity and destination choices for the whole population. Finally, evacuation of over a million residents in the study area was simulated using agent-based simulation built in MATSim. The agent-based modeling framework proposed in this paper provides an integrated methodology for evacuation simulation with specific consideration of agents' behaviors. The simulation results need to be further validated and verified using real-world evacuation data.

AB - Transportation modeling and simulation play an important role in the planning and management of emergency evacuation. It is often indispensable for the preparedness and timely response to extreme events occurring in highly populated areas. Reliable and robust agent-based evacuation models are of great importance to support evacuation decision making. Nevertheless, these models rely on numerous hypothetical causal relationships between the evacuation behavior and a variety of factors including socio-economic characteristics and storm intensity. Understanding the impacts of these factors on evacuation behaviors (e.g., destination and route choices) is crucial in preparing optimal evacuation plans. This paper aims to contribute to the literature by integrating well-calibrated behavior models with an agent-based evacuation simulation model in the context of hurricane evacuation. Specifically, discrete choice models were developed to estimate the evacuation behaviors based on large-scale survey data in Northern New Jersey. Monte-Carlo Markov Chain (MCMC) sampling method was used to estimate evacuation propensity and destination choices for the whole population. Finally, evacuation of over a million residents in the study area was simulated using agent-based simulation built in MATSim. The agent-based modeling framework proposed in this paper provides an integrated methodology for evacuation simulation with specific consideration of agents' behaviors. The simulation results need to be further validated and verified using real-world evacuation data.

KW - agent-based simulation

KW - demand modeling

KW - Hurricane evacuation

KW - MATSim

UR - http://www.scopus.com/inward/record.url?scp=85051263570&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85051263570&partnerID=8YFLogxK

U2 - 10.1016/j.procs.2018.04.074

DO - 10.1016/j.procs.2018.04.074

M3 - Conference article

VL - 130

SP - 836

EP - 843

JO - Procedia Computer Science

JF - Procedia Computer Science

SN - 1877-0509

ER -