Multi-Scale Evacuation Models To Support Emergency And Disaster Response

Evacuation is a short-term measure to mitigate human injuries and losses by temporarily relocation of exposed population before, during, or after disasters. With the increasing growth of population and cities, buildings and urban areas are over-populated which brings about safety issues when there is a need for emergency evacuation. In disaster studies, simulation is widely used to explore how natural hazards might evolve in the future, and how societies might respond to these events. Accordingly, evacuation simulation is a potentially helpful tool for emergency responders and policy makers to evaluate the required time for evacuation and the estimated number and distribution of casualties under a disaster scenario. The healthcare system is an essential subsystem of communities which ensures the health and well-being of their residents. Hence, the resilience of the healthcare system plays an essential role in the resilience of the whole community. In disasters, patient mobility is a major challenge for healthcare systems to overcome. This is where the scientific society enters with modeling and simulation techniques to help decision-makers. Hospital evacuation simulation considering patients with different mobility characteristics, needs, and interactions, demands a microscopic modeling approach, like Agent-Based Modeling (ABM). However, as the system increases in size, the models become highly complex and intractable. Large-scale complex ABMs can be reduced by reformulating the micro-scale model of agents by a meso-scale model of population densities and partial differential equations, or a macro-scale model of population stocks and ordinary differential equations. However, reducing the size and fidelity of microscopic models to meso- or macro-scale models implies certain drawbacks. This dissertation contributes to the improvement of large-scale agent-based evacuation simulation and multi-scale hospital evacuation models. For large-scale agent-based models, application of bug navigation algorithms, popular in the field of robotics, is evaluated to improve the efficiency of such models. A candidate bug algorithm is proposed based on a performance evaluation framework, and its applicability and practicability are demonstrated by a real-world example. For hospital evacuation simulation, crowd evacuation considering people with different physical and mobility characteristics is modeled on three different scales: microscopic (ABM), mesoscopic (fluid dynamics model), and macroscopic (system dynamics model). Similar to the well-known Predator-Prey model, the results of this study show the extent to which macroscopic and mesoscopic models can produce global behaviors emerging from agents’ interactions in ABMs. To evaluate the performance of these multi-scale models, the evacuation of the emergency department at Johns Hopkins University is simulated, and the outputs and performance of the models are compared in terms of implementation complexity, required input data, provided output data, and computation time. It is concluded that the microscopic agent-based model is recommended to hospital emergency planners for long-term use such as evaluating different emergency scenarios and effectiveness of different evacuation plans. On the other hand, the macroscopic system dynamics model is best to be used as a simple tool (like an app) for rapid situation assessment and decision making in case of imminent events. The fluid dynamics model is found to be suitable only for studying crowd dynamics in medium to high densities, but it does not offer any competency as an evacuation simulation tool