Scalable Epidemiological Workflows to Support COVID-19 Planning and Response

Abstract

The COVID-19 global outbreak represents the most significant epidemic event since the 1918 influenza pandemic. Simulations have played a crucial role in supporting COVID-19 planning and response efforts. Developing scalable workflows to provide policymakers quick responses to important questions pertaining to logistics, resource allocation, epidemic forecasts and intervention analysis remains a challenging computational problem. In this work, we present scalable high performance computing-enabled workflows for COVID-19 pandemic planning and response. The scalability of our methodology allows us to run fine-grained simulations daily, and to generate county-level forecasts and other counterfactual analysis for each of the 50 states (and DC), 3140 counties across the USA. Our workflows use a hybrid cloud/cluster system utilizing a combination of local and remote cluster computing facilities, and using over 20,000 CPU cores running for 6-9 hours every day to meet this objective. Our state (Virginia), state hospital network, our university, the DOD and the CDC use our models to guide their COVID-19 planning and response efforts. We began executing these pipelines March 25, 2020, and have delivered and briefed weekly updates to these stakeholders for over 30 weeks without interruption.

Deploying Vaccine Distribution Sites for Improved Accessibility and Equity to Support Pandemic Response

“…Data: We run our experiments on the mobility data from Charlottesville City and Albemarle County in Virginia. For these counties, we use synthetic data constructed from the 2019 U.S. population pipeline (see [33,34] for details). This dataset was constructed by tracking the week-long activity of county residents.…”

COVID’s collateral damage: likelihood of measles resurgence in the United States

“…The social contact network of the Virginia population consists of more than 7.6 million nodes and 371.9 million edges, and was developed using a "first principles" approach [12,13,15,16]. The nodes represent synthetic individuals with respective households which are located geographically.…”

“…We use a tool called EpiHiper [16] to simulate the spread of measles under various scenarios over the social network of Virginia. It is implemented in C + + /MPI and is scalable to millions of agents via its parallel algorithm, which enables scaling on distributed memory systems [16].…”

“…We use a tool called EpiHiper [16] to simulate the spread of measles under various scenarios over the social network of Virginia. It is implemented in C + + /MPI and is scalable to millions of agents via its parallel algorithm, which enables scaling on distributed memory systems [16]. EpiHiper is a network-based epidemiological model designed to study the impact of individual behavior and public health policies on the spread of infectious diseases.…”

Data-driven scalable pipeline using national agent-based models for real-time pandemic response and decision support

“…We implemented a new agent-based model (EpiHiper) and developed digital twin populations to support the COVID-19 pandemic response. Case studies done in the context of COVID-19 response, using earlier versions of the modeling environment include: ( i ) network-based vaccine allocation ( Chen et al, 2021 , 2022 ); ( ii ) contact tracing ( Hoops et al, 2021 ); ( iii ) economic impact analysis ( Chen et al, 2020a , b ); and ( iv ) the role of basic non-pharmaceutical interventions ( Machi et al, 2021 ).…”