An Optimization Framework to Study the Balance Between Expected Fatalities due to COVID-19 and the Reopening of U.S. Communities

Abstract

As communities reopen following shelter-in-place orders, they are facing two conflicting objectives. The first is to keep the COVID-19 fatality rate down. The second is to revive the U.S. economy and the livelihood of millions of Americans. In this paper, a team of researchers from the Center on Stochastic Modeling, Optimization, & Statistics (COSMOS) at the University of Texas at Arlington, in collaboration with researchers from University of Texas Southwestern Medical Center and Harvard Medical School, has formulated a computationally-efficient optimization framework, referred to as COSMOS COVID-19 Linear Programming (CC19LP), to study the delicate balance between the expected fatality rate and the level of normalcy in the community. Given the disproportionate fatality characteristics of COVID-19 among those in different age groups or with an underlying medical condition or those living with crowding, the key to the CC19LP framework is a focus on "key contacts" that separate individuals at higher risk from the rest of the population. The philosophy of CC19LP lies in maximizing protection of key contacts, so as to shield high-risk individuals from infection. Given the lack of pharmaceutical solutions, i.e., a vaccine or cure, the CC19LP framework minimizes expected fatalities by optimizing the use of non-pharmaceutical interventions, namely COVID-19 testing; personal protective equipment; and social precautions, such as distancing, hand-washing, and face coverings. Low-risk individuals that are not key contacts, including most children, are unrestricted and can choose to participate in pre-pandemic normal activities, which eliminates the need for compliance across the entire population. Consequently, the CC19LP framework demonstrates optimal strategies for protecting high-risk individuals while reopening communities.

Optimizing Non-Pharmaceutical Intervention Strategies Against COVID-19 Using Artificial Intelligence

“…However, the optimization did not operate on real-life NPIs, and as such, this approach cannot be directly used by policy-makers. Chen et al 8 created a linear programming tool to explore the trade-off between the expected mortality rate of COVID-19 and return to normal activities, while Yaesoubi et al 9 developed a decision tool to determine when to trigger, continue, or stop physical distancing intervention in order to minimize both the deaths from COVID-19 and intervention duration. Both studies combined the objectives into a single function and the final result was a single intervention plan.…”

Optimization in the Context of COVID-19 Prediction and Control: A Literature Review

“…Stochastic optimization (e.g., [167]) and game theory approaches (e.g., [168]) were implemented to determine the optimal timing and duration of social distancing policies, and other mathematical programming models were developed to determine both personal and mass protection strategies [169], as well as targeted immunization models [170]. In [171], the authors created a linear programming model to study the trade-off between the expected fatality rate due to COVID-19 and the return to normal activities. Deaths were minimized by optimizing the implementation of nonpharmaceutical interventions, such as social distancing and mask mandates.…”

Optimizing non-pharmaceutical intervention strategies against COVID-19 using artificial intelligence

“…However, the optimization did not operate on real-life NPIs, and as such, this approach cannot be directly used by policy-makers. Chen et al ( 7 ) created a linear programming tool to explore the trade-off between the expected mortality rate of COVID-19 and return to normal activities, while Yaesoubi et al ( 8 ) developed a decision tool to determine when to trigger, continue, or stop physical distancing intervention in order to minimize both the deaths from COVID-19 and intervention duration. Both studies combined the objectives into a single function and the final result was a single intervention plan.…”