Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Finding critical transitions of the post-disaster recovery using the sensitivity analysis of agent-based models

Sangung Park, Jiawei Xue, Satish V. Ukkusuri

TL;DR

This paper develops an ABM-MN framework to study post-disaster recovery, introducing threshold-based return decisions and a data-driven SD model within a multilayer network to capture cross-layer interactions. By comparing threshold models against a binary logit, it demonstrates that threshold rules can replicate key macroscopic recovery patterns while revealing conditions under which CTs emerge, notably high thresholds and small populations (< $1{,}000$). The study applies the framework to Hurricane Harvey-impacted counties to uncover urban-rural differences in recovery trajectories and emphasizes the importance of threshold choice and population size for robust PDR modeling. Practically, the work provides policy-relevant insights into the risk of abrupt transitions in rural areas and offers a data-informed path to improve recovery planning and resilience assessments.

Abstract

Frequent and intensive disasters make the repeated and uncertain post-disaster recovery process. Despite the importance of the successful recovery process, previous simulation studies on the post-disaster recovery process did not explore the sufficient number of household return decision model types, population sizes, and the corresponding critical transition conditions of the system. This paper simulates the recovery process in the agent-based model with multilayer networks to reveal the impact of household return decision model types and population sizes in a toy network. After that, this paper applies the agent-based model to the five selected counties affected by Hurricane Harvey in 2017 to check the urban-rural recovery differences by types of household return decision models. The agent-based model yields three conclusions. First, the threshold model can successfully substitute the binary logit model. Second, high thresholds and less than 1,000 populations perturb the recovery process, yielding critical transitions during the recovery process. Third, this study checks the urban-rural recovery value differences by different decision model types. This study highlights the importance of the threshold models and population sizes to check the critical transitions and urban-rural differences in the recovery process.

Finding critical transitions of the post-disaster recovery using the sensitivity analysis of agent-based models

TL;DR

This paper develops an ABM-MN framework to study post-disaster recovery, introducing threshold-based return decisions and a data-driven SD model within a multilayer network to capture cross-layer interactions. By comparing threshold models against a binary logit, it demonstrates that threshold rules can replicate key macroscopic recovery patterns while revealing conditions under which CTs emerge, notably high thresholds and small populations (< ). The study applies the framework to Hurricane Harvey-impacted counties to uncover urban-rural differences in recovery trajectories and emphasizes the importance of threshold choice and population size for robust PDR modeling. Practically, the work provides policy-relevant insights into the risk of abrupt transitions in rural areas and offers a data-informed path to improve recovery planning and resilience assessments.

Abstract

Frequent and intensive disasters make the repeated and uncertain post-disaster recovery process. Despite the importance of the successful recovery process, previous simulation studies on the post-disaster recovery process did not explore the sufficient number of household return decision model types, population sizes, and the corresponding critical transition conditions of the system. This paper simulates the recovery process in the agent-based model with multilayer networks to reveal the impact of household return decision model types and population sizes in a toy network. After that, this paper applies the agent-based model to the five selected counties affected by Hurricane Harvey in 2017 to check the urban-rural recovery differences by types of household return decision models. The agent-based model yields three conclusions. First, the threshold model can successfully substitute the binary logit model. Second, high thresholds and less than 1,000 populations perturb the recovery process, yielding critical transitions during the recovery process. Third, this study checks the urban-rural recovery value differences by different decision model types. This study highlights the importance of the threshold models and population sizes to check the critical transitions and urban-rural differences in the recovery process.
Paper Structure (28 sections, 7 equations, 5 figures, 5 tables)

This paper contains 28 sections, 7 equations, 5 figures, 5 tables.

Table of Contents

  1. Abstract
  2. Introduction
  3. Data descriptions
  4. Hurricane Harvey and the five counties
  5. Mobile phone location data
  6. POI data
  7. Post-hurricane survey data
  8. The toy network
  9. An ABM-MN simulation framework
  10. Definitions of MN and agents
  11. The data-driven SD model in the ABM-MN
  12. The agent's return decision models
  13. The binary logit model
  14. The threshold model
  15. Updating rules of ABM-MN
  16. ...and 13 more sections

Figures (5)

  • Figure 1: Schematic overview of this study. (a) Schematic illustration of a multilayer network and multi-agent systems (MAS). (b) Types of MAS interactions and focuses of this study. (c) Spatial and temporal scales of this study. A toy network consisting of MNs (Top). A trajectory of Hurricane Harvey in Texas, 2017 (Bottom).
  • Figure 2: Simulation results over two months by different types of human interactions and population sizes after September 1, 2017. (a) Simulation results aggregated by model types. The blue line shows the logit model results. The orange, green, red, and purple lines show a value of 0.6, 0.7, 0.8, and 0.9 of the universally homogeneous threshold. The dark purple line delineates a universally heterogeneous threshold model (hetero). The pink line represents an individually heterogeneous threshold model (different). The grey line visualizes a universally time-varying threshold model (time-varying). (b) and (c) Examples of simulation results by the BLM and the homogeneous threshold model (0.6). The blue line represents the human values when the population ranges from 100 to 1000. The other colored lines represent the human values ranging from the first value to the last value.
  • Figure 3: Sensitivity analysis of the ABM-MN. (a) The average of human value residuals by populations and thresholds. (b) The sum of residuals by different types of interactions.
  • Figure 4: Real-world simulation results. (a) Average human values over two months by the model types in five selected counties in Hurricane Harvey. (b) Visualization of human value on five counties at day 0. (c) Visualization of human value on five counties on the day 30.
  • Figure 5: Urban-Rural differences in the simulation results. (a) Human value over time in Harris County by agent's return decision model types. (b) Human value over time in Brazoria County by agent's return decision model types. (c) The human value difference between Harris County and Brazoria County.