Epidemiologically and Socio-economically Optimal Policies via Bayesian Optimization

Abstract

Mass public quarantining, colloquially known as a lock-down, is a non-pharmaceutical intervention to check spread of disease. This paper presents ESOP (Epidemiologically and Socioeconomically Optimal Policies), a novel application of active machine learning techniques using Bayesian optimization, that interacts with an epidemiological model to arrive at lock-down schedules that optimally balance public health benefits and socioeconomic downsides of reduced economic activity during lock-down periods. The utility of ESOP is demonstrated using case studies with VIPER (Virus-Individual-Policy-EnviRonment), a stochastic agent-based simulator that this paper also proposes. However, ESOP is flexible enough to interact with arbitrary epidemiological simulators in a black-box manner, and produce schedules that involve multiple phases of lock-downs. Keywords Optimal policy • Lock-down • Epidemiology • Bayesian optimization Disclaimer This paper makes no recommendation to individuals and its results should not be interpreted by individuals to modulate personal behavior. The authors recommend that individuals continue to follow guidelines offered by local governments with respect to lock-downs and social distancing, and those offered by medical professionals with respect to personal hygiene and treatment.

EpidemiOptim: A Toolbox for the Optimization of Control Policies in Epidemiological Models

“…(Kompella et al, 2020) applied the Soft-Actor-Critic algorithm (Haarnoja et al, 2018) on an agent-based epidemiological model with community interactions allowing the spread of the disease to be an emergent property of people's behaviors and the government's policies. Other contributions applied non-rl optimization methods such as deterministic rules (Tarrataca et al, 2020), stochastic approximation algorithms (Yaesoubi et al, 2020), optimal control (Charpentier et al, 2020) or Bayesian optimization (Chandak et al, 2020). This latter paper also proposes a stochastic agent-based model called VIPER (Virus-Individual-Policy-EnviRonment) allowing to compare the optimization results on variations of the demographics and geographical distribution of population.…”

On the interplay between mobility and hospitalization capacity during the COVID-19 pandemic: The SEIRHUD model

“…The SEIR epidemiological model, and several variations of it, have been extensively studied both for the purpose of understanding dynamics [10,11,12,8,13,14,15] as well as optimal policy making [16,17,18,19,20]. For a comprehensive review see [21,22].…”

PREFACE on the Special Issue ‘Technologies for Fighting COVID-19’

“…Ranjan ( 2020 ) proposed two data-driven models to forecast the decay phase of the epidemic, as the epidemic in the decay phase is different from its growth phase. Chandak et al ( 2020 ) described a machine learning based application to compute the lock-down schedules, while taking health and economic related activities into consideration. Khadilkar et al ( 2020 ) examined the lock-down policies, using an AI-driven approach, which can simultaneously control the spread of the disease while balancing it with both health and economic costs.…”