Agent-Based Simulation of Building Evacuation after an Earthquake: Coupling Human Behavior with Structural Response

Z. Liu, C. C. Jacques, S. Szyniszewski, J. K. Guest, B. W. Schafer, Takeru Igusa, Judith Mitrani-Reiser

Research output: Contribution to journalArticle

Abstract

The safety of building occupants during and immediately after disasters, such as a major earthquake, is highly dependent on the way in which people interact with the damaged physical environment. While there are extensive studies on evacuation from undamaged structures and on structural behavior under seismic and other hazards, research on the influence of building damage on human evacuation behavior is limited. This study presents a framework by which models for buildings and human behavior can be coupled to analyze the dynamic influences of building damage on the evacuation process. The framework combines nonlinear dynamic finite-element modeling of structures, probabilistic modeling of damage, and agent-based modeling of human occupants to investigate the behavior of people as they interact with each other and with their dynamically-deteriorating environment as they attempt to evacuate the building. A case study is presented for a typical three-story commercial office building subjected to the ground motions of the 1994 Northridge, California earthquake. By using exit flow rates and other measures related to evacuation time histories as the outcomes of interest, it is shown how the proposed framework can be used as a tool to enhance building design and to develop recommendations for improved evacuation strategies. An important future extension of the work is expanding the framework for multiple buildings for community-wide models of postdisaster behavior.

Original languageEnglish (US)
Article number04015019
JournalNatural Hazards Review
Volume17
Issue number1
DOIs
Publication statusPublished - Feb 1 2016

    Fingerprint

ASJC Scopus subject areas

  • Civil and Structural Engineering
  • Environmental Science(all)
  • Social Sciences(all)

Cite this