Modeling and Simulating the Network Behavior of Agencies during Disaster Relief Operations
Coles, John B.
MetadataShow full item record
The objective of this dissertation is to advance the study of disaster response and recovery (generally, disaster relief) by providing insight and tools to agencies that work in disaster relief. In disaster relief, resources such as time, money, and personnel are essential to the success of the effort and must be closely guarded and wisely invested. As the scale of the disaster increases (e.g., amount of need), so does the number of agencies, the amount of money, and the number of volunteers. As people and resources converge onto a community impacted by a disaster, the problems of partnership management and coordination become increasingly complicated. This dissertation explores some of the common challenges of interagency coordination and looks at how game theory, network analysis, and agent-based simulation/optimization can provide insight and solutions for people and agencies working to help disaster survivors. This dissertation was developed in response to countless hours working alongside and talking with emergency managers and aid workers who wanted tools that could help them determine with greater accuracy the amount of resources and types of partnerships that are worth investing in following a disaster or extreme event. The recent earthquakes in Japan, Chili, China, and Haiti, as well as the ongoing preparation for similar earthquakes in the New Madrid Seismic Zone in the U.S., make it imperative to increase our understanding of interagency dynamics after a disaster. This dissertation explores the development of a model that will support statistical analysis of agency behavior in a network during response operations around the world. Some of the key advances presented in the dissertation are: (a) An optimization model that explores the concept of resource allocation in a disaster relief environment using the International Federation of Red Cross and Red Crescent Society's Code of Conduct; (b) A game theory-based model for partner selection in the presence of uncertainty; and (c) An interagency network simulation engine called DRAMAS (Disaster Response Agent-based network Management and Adaptation System) developed based on current disaster relief literature, hands-on experience, and the results of interviews and experiments in disaster operations. Preliminary results of this work has been presented at conferences and published in peer-reviewed journals over the last five years. Research partnerships with disaster relief organizations around the world have yielded high quality data about decision making in disaster environments, and have proved to be an excellent avenue for the dissemination of research findings. Data from interviews and experiments have provided the mechanics of DRAMAS for how agencies interact during disaster relief operations, and is designed to be scaled to a variety of different scenarios. Work in partnership selection and behavior has yielded the publication of three papers in peer-reviewed journals, several working papers on data collected after disasters, and insight into the mechanics of the interagency partner selection and resource allocation process. Finally, work in the area of network behavior has provided interesting insights and new tools for data collection, analysis, and dissemination. This work is built on extensive research of disaster management and provides a comprehensive analysis of agency posturing following an extreme event. The DRAMAS model uses game theory and stochastic processes to explore the complex interactions between relief agencies of different sizes and capabilities. The decision models in Chapters 3-5, and the DRAMAS simulation environment provide an excellent testing ground for hypotheses regarding relief agency partnerships, goals, roles, and prior involvement, by providing a depiction of the change in agency partnerships and resource investments following a disaster. The goal of this research is to expand the current body of knowledge and examine the fundamental principles of agency success during relief operations. Results from this work provide a path for improving our understanding of interagency partnerships and interaction, and could provide new insights into the behavior of agency networks in response to a disaster.