Agent-based simulation of seismic crisis including human behavior: application to the city of Beirut, Lebanon

Abstract

Earthquake simulations at the urban scale usually focus on estimating the damages to the built environment and the consequent losses without fully taking into account human behavior in crisis. Yet, human behavior is a key element for improving crisis disaster management; therefore, it is important to include it in seismic crisis simulations. In this study, an agent-based model for the simulation of pedestrian evacuation during earthquakes at the city scale is developed following an interdisciplinary approach. The model recreates the urban conditions using Geographic Information System (GIS) and a synthetic population, in addition to the earthquake consequences on the urban fabric. Moreover, the model integrates realistic human behaviors calibrated using quantitative survey results. We simulate pedestrian outdoor mobility with the different constraints that affect it such as the topography and the presence of debris. The simulator is applied to the case of Beirut, Lebanon. A what-if approach is adopted to analyze the population’s safety in case of earthquakes in Beirut, particularly the open spaces’ capacity to provide shelters and the effect of debris and realistic human behaviors on people’s safety. The simulation results show that less than 40% of the population is able to arrive at an open space within 15 min after an earthquake. This number is further reduced when some open spaces are locked. Debris and realistic human behaviors significantly delay the arrivals to safe areas and, therefore, should not be neglected in earthquake simulations.

A data-driven approach of layout evaluation for electric vehicle charging infrastructure using agent-based simulation and GIS

“…For instance, Davidson et al 48 developed a discrete-event spatial model to simulate disease spread within a pandemic, demonstrating its adaptability and ability to provide deterministic predictions for multiple regions simultaneously. Other studies using simulation to investigate the spread of infectious diseases can be found in the work by Abadeer et al, 49 as well as the review article by Ayadi et al 50 Iskandar et al 51 developed an agent-based model that incorporates realistic human behavior and urban conditions to simulate pedestrian evacuation during earthquakes at the city scale, revealing the impact of debris and human behavior on population safety and the limited capacity of open spaces to provide shelters in Beirut, Lebanon. Risco-Martin et al 52 introduced a discrete-event simulation for real-time monitoring and management of harmful algal and cyanobacterial blooms (HABs), addressing the need for efficient detection and response to HABs threatening water quality in dynamic environments.…”