Disaster Response Applications Using Agent-Based Modeling
Battelle is the world’s largest, non-profit, independent R&D organization, and is a worldwide leader in the development, commercialization, and transfer of technology. They manage or co-manage laboratories for the U.S. Department of Energy, the U.S. Department of Homeland Security, and an international nuclear laboratory in the United Kingdom.
In an effort to find practical operational solutions for a fast and effective response to an unexpected crisis or natural disaster, Battelle needed to test the effectiveness of a 48 hour shelter-in-place order for an Improvised Nuclear Device scenario (IND). The intended goal was to reduce radiation dosages received during an uncoordinated mass evacuation, by comparing immediate evacuation and shelter-in-place order.
Modeling a disaster, whether natural or man-made, represents many unique challenges. There are distinctive environments and physical consequences, and numerous scenario possibilities and threat vectors. In addition, response strategies are rarely implemented as planned, and there are unknown human reactions.
Simulation was chosen for the disaster modeling because it had the capability to evaluate the space of potential scenarios. Deterministic models had limitations incorporating factors, like fundamentally unpredictable human responses and the need to compare alternatives versus looking for exact answers.
AnyLogic software was a natural choice for Battelle, as the software was already being utilized in a broad range of projects within the organization, including:
In addition, AnyLogic’s agent-based capabilities allowed Battelle to capture the most important dynamics of a disaster event. Emergence, or emergent behavior, is a key principal in modeling human behavior. Also, a model can sometimes exhibit unexpected outcomes. Both of these issues can only be captured using agent-based modeling.
Disaster Response Model Framework
The comprehensive model framework included an environment of road networks, vehicles, drivers, and disaster events. The road network was built with road layouts from GIS databases, local highway agency data (speed limits, lane capacity), and agents as node points for greater control. Changes to the network, such as the flooding of roads or destruction of bridges, were incorporated into dynamic events as the disaster unfolded.
The physical limitations of vehicles were governed by parameter data provided by the US Census, Bureau of Transportation. Data from past disaster response studies was used to represent driver agent behaviors, taking into account the changes in irrational drivers in normal circumstances versus during a mass evacuation. The model also incorporated dynamic route finding (several interlinked agent state sets that were dynamically tracked and updated). In addition, all behavior states were linked to physical vehicle movement parameters to initiate vehicle stoppages as drivers became incapacitated.
Agent behavior variables from initial values were calibrated, and evacuation data was used from past disasters to set accuracy targets, since calibration and validation were critical steps in proving the validity of the simulation model. If no historical data was available, Battelle used data from other major transportation events, sensitivity analysis based on other disaster events, and survey data.
Dynamic contours were used to track regions of disaster consequences, often derived from other simulation models, to compartmentalize processing requirements. Contours updated in real time based on predicted weather patterns, land cover, etc., and multiple interlinked contour sets could be adapted to represent almost any disaster scenario (for example, flooding levels, fire spread, damage path, contamination/fallout spread). In the IND scenario, two main contour sets were used; blast radius levels (fireball and overpressure force contours) and fallout distribution (radiation levels in air and deposition on ground from various radioactive particle types).
The simulation model built using AnyLogic software compared immediate evacuation versus shelter-in-place order and showed that shelter-in-place order significantly reduced radiation dosage received, as well as cases of severe radiation poisoning for large INDs.
The model also produced downstream model outputs to test different disaster response strategies and find the best response strategy among several likely options. Battelle was able to incorporate emergency responder agents, multiple intervention scenarios, and interchangeable model components (different locations for same disaster scenario, or different scenario for same location), to achieve the goal of finding practical operational solutions for fast and effective responses to various unexpected crisises or natural disasters.
More Case Studies
Oil Pipeline Network Development: Finding Bottlenecks and Choosing the Right PoliciesOne of the largest oil and gas pipeline operators in North America was delivering oil to a client that was not always able to accept the incoming batches. The operator was challenged to quantify the system impacts of deferred downstream deliveries. They also needed to determine whether the existing tankage at upstream oil terminals would be adequate to store the deferred batches.
Simulation of Maternity Ward OperationsThis model simulates the maternity ward in a hospital currently under construction. The purpose of the model is to support discussions related to which resources, capacity, and work methods are required on the new ward. The project was carried out for Karolinska University Hospital in the Stockholm County, Sweden.
Evaluating Hospital Inpatient Care CapacityStockholm County, Sweden was in the process of building a new, highly specialized hospital. The Health Administration of the county questioned whether they would get an acceptable level of care production with the current investments and reasoning concerning various operational and strategic issues. To find the answers, they used simulation modeling in AnyLogic.
Handling Total Care Need for Dialysis PatientsThe County of Stockholm (Sweden), like any country or region, experiences a continuous need to handle the healthcare necessities of various patient groups. Each group can be seen as a subpopulation, with its own distinctive traits, characteristics, and challenges. The discussed simulation project focused on the dialysis patients, a group who needs to visit caregiving facilities frequently.
Evaluating Healthcare Policies to Reduce Rates of Cesarean DeliveryThe challenge of reducing the cesarean delivery rate has been recognized by numerous researchers for years. For the first time, in research conducted for the Washington State, Alan Mills, FSA MAAA ND, a research actuary, and his colleagues reproduced this part of the United States healthcare system in a simulation model to allow the stakeholders, including health agencies, insurers, clinicians, and legislators, to test their assumptions on the model to find the right solutions.
Shaping Healthcare Policy Using SimulationAn initiative by the Department of Mechanical and Industrial Engineering at the University of Toronto, the Centre for Research in Healthcare Engineering (CRHE), was in response to the immediate and compelling desire for efficiency and quality improvements in the Canadian healthcare system.
An Agent-Based Explanation for SPMI Living Situation ChangesOver the past 60 years, the number of Severely and Persistently Mentally Ill (SPMI) patients in the US living in the community increased. Yet a growing minority of people with severe illness are worse off because they are homeless or incarcerated. In this case study, IBM Global Research and Otsuka Pharmaceuticals used an agent-based approach to model these remarkable swings.
Outpatient Appointment Scheduling Using Discrete Event Simulation ModelingIndiana University Health Arnett Hospital, consisting of a full-service acute care hospital and a multispecialty clinic, faced poor statistics because the number of no-show patients (those who don’t show up for their scheduled appointments) rose dramatically to 30%. This was primarily connected to the fact that clinic schedules were driven by individual preferences of the medical staff, which led to increased variations in scheduling rules. To eliminate the problem, the client wanted to develop a scheduling methodology that would benefit the clinic, doctors, and patients.