Collaborative Research: Simulation-Based Policy Analysis for Reducing Ebola Transmission Risk in Air Travel

Project: Research project

Description

Air travel has been identified as a leading factor in the spread of infections, such as influenza, and this has motivated calls for limitations on air travel in order to prevent the spread of Ebola. However, such limitations carry considerable economic and human costs. Consequently, it is necessary to evaluate the extent of impact of air travel on spread of Ebola and to also identify policy options that can mitigate its spread without major disruption to air travel. Given the political pressures for taking action, there is an urgent need to provide tools that will enable policy makers to take rational decisions in this matter. Our goal is to use our teams expertise in modeling human movement in planes, modeling the spread of infections, software infrastructure for decision support, and massively parallel computing to provide useful insight to decision makers.

We need fine-scale causal models to provide insight into the consequences of different policy choices. Inherent uncertainties in human behavior make precise prediction difficult. This is further magnified by the practical reality that models for all factors influencing an epidemic are not even available, especially in the initial stages. Such sources of uncertainty are parameterized and our goal is to identify potential vulnerabilities in different policy options across the parameter space.

Our model for human movement is motivated by the idea of molecular dynamics. We define potential fields around individuals and movement is directed by the interaction of these potentials, with constraints based on physical barriers, and driving forces, such as boarding, etc. Our prior results model evacuation of a plane with 5% error. We will modify the parameters of the potential to deal with different situations, such as boarding, in-flight movement, and disembarkation. We will integrate this with Ebola transmission model to determine potential spread of Ebola under different scenarios for 15 different aircraft types, which account for a significant fraction of passenger-miles flown. We will perform phylogeographic analysis using existing genetic data for new and old strains to identify the spread of Ebola, and integrate these with our air-transport model. This integration will provide mutual feedback to the models, enabling better calibration of human movement parameters and better understanding of the potential for spread of Ebola through air-transport.

In order to enable the above results to be useful in evaluating policy decisions, we will integrate them with CSF. This work will provide insight into suitable interfaces required for models dealing with factors influencing epidemic mitigation policy. We will scale CSF to the petascale using new techniques we have developed for topology and routing aware mapping of tasks to nodes on massively parallel machines. We will also reduce the number of scenarios evaluated in the parameter space by dynamically generating finer resolution in regions of parameter space with rapid changes in outcomes and fewer in less sensitive regions. This will trade-off slightly higher communication cost for substantial potential reduction in computational cost.
The impact of this work will go beyond the urgent policy questions we answer. This research will also lay the foundation for a scalable software infrastructure that can identify good policies for new infectious diseases at their onset, when detailed knowledge is sparse. It will also provide a new class of applications targeting future generation supercomputers.

StatusFinished
Effective start/end date4/1/1512/31/16

Funding

  • National Science Foundation (NSF): $100,000.00

Fingerprint

Air
Costs
Supercomputers
Parallel processing systems
Molecular dynamics
Aircraft
Topology
Calibration
Feedback
Economics
Communication
Uncertainty