Feedback Design for Devising Optimal Epidemic Control Policies

TL;DR

This paper introduces a feedback-based approach for robust epidemic monitoring and control that accounts for data noise and model uncertainties, using observer design and optimal control on a modified SIR model.

Contribution

It presents a novel observer design for robust state estimation and optimal control policy formulation under uncertainty in epidemic models.

Findings

Effective in mitigating uncertainties in epidemic monitoring.

Robust control policies improve epidemic management.

Demonstrated on a modified SIR model.

Abstract

This paper proposes a feedback design that effectively copes with uncertainties for reliable epidemic monitoring and control. There are several optimization-based methods to estimate the parameters of an epidemic model by utilizing past reported data. However, due to the possibility of noise in the data, the estimated parameters may not be accurate, thereby exacerbating the model uncertainty. To address this issue, we provide an observer design that enables robust state estimation of epidemic processes, even in the presence of uncertain models and noisy measurements. Using the estimated model and state, we then devise optimal control policies by minimizing a predicted cost functional. To demonstrate the effectiveness of our approach, we implement it on a modified SIR epidemic model. The results show that our proposed method is efficient in mitigating the uncertainties that may arise in epidemic monitoring and control.