Register now After registration you will be able to apply for this opportunity online.
This opportunity is not published. No applications will be accepted.
Optimal Adaptive design of Non-Pharmaceutical Interventions for Network Epidemics
Understanding how to effectively control an epidemic spreading on a network is a problem ofparamount importance for the scientific community. The ongoing COVID-19 pandemic has high-lighted the need for policies that mitigate the spread, without relying on pharmaceutical interven-tions, that is, without the medical assurance of the recovery process. These policies typically entaillock-downs and mobility restrictions, having thus non negligible socio-economic consequences forthe population. In this work, we focus on the problem of finding the optimum policies that “flattenthe epidemic curve” while limiting the negative consequences for the society. To do so, we want to implement an Adaptive Model Predictive Control (MPC) scheme where the parameters of the epidemics model are learned from experimental data.
The ongoing COVID-19 pandemic has highlighted the key role played by public health authorities in enacting Non Pharmaceutical Interventions (NPI) to "flatten the epidemic curve" when no effective pharmaceutical treatments such as vaccines are available. However, NPIs typically entail the implementation of harsh measures, including lockdowns and restrictions of personal freedom of movement, which may yield severe socio-psychological and economic consequences. Thus, they should be implemented keeping a reasonable balance between safety and normalcy. To this aim, the development of tools to predict the course of an epidemic and evaluate the impact of different NPIs has become a task of paramount importance for the scientific community, aiming at assisting public health authorities in their decisions.
The mathematical modeling of epidemics has emerged as a valuable framework to perform such a task. Relevant examples can be found in the useful insights provided into the ongoing COVID-19 pandemic and more in general on the spread of epidemics in over a network of communities, viz. states or cities.
Motivated by the encouraging predictive abilities of these models and by the recent developments in control schemes for optimal NPIs design, we focus in this work on the development of a design based on Adaptive MPC. The strength of this approach is the capability of learning online the model coefficients making it robust to changes in the virus or in the environment, and consequently reliable for the policy-maker that has to finally select the NPIs.
The ongoing COVID-19 pandemic has highlighted the key role played by public health authorities in enacting Non Pharmaceutical Interventions (NPI) to "flatten the epidemic curve" when no effective pharmaceutical treatments such as vaccines are available. However, NPIs typically entail the implementation of harsh measures, including lockdowns and restrictions of personal freedom of movement, which may yield severe socio-psychological and economic consequences. Thus, they should be implemented keeping a reasonable balance between safety and normalcy. To this aim, the development of tools to predict the course of an epidemic and evaluate the impact of different NPIs has become a task of paramount importance for the scientific community, aiming at assisting public health authorities in their decisions.
The mathematical modeling of epidemics has emerged as a valuable framework to perform such a task. Relevant examples can be found in the useful insights provided into the ongoing COVID-19 pandemic and more in general on the spread of epidemics in over a network of communities, viz. states or cities. Motivated by the encouraging predictive abilities of these models and by the recent developments in control schemes for optimal NPIs design, we focus in this work on the development of a design based on Adaptive MPC. The strength of this approach is the capability of learning online the model coefficients making it robust to changes in the virus or in the environment, and consequently reliable for the policy-maker that has to finally select the NPIs.
- Learn about (network) epidemic models for COVID-19;
- Formalize the problem of optimal design of NPIs for the selected epidemic model;
- Develop an Adaptive MPC scheme for the problem;
- Validate your algorithm via numerical simulations based on real data obtained from the current COVID-19 pandemic.
- Learn about (network) epidemic models for COVID-19; - Formalize the problem of optimal design of NPIs for the selected epidemic model; - Develop an Adaptive MPC scheme for the problem; - Validate your algorithm via numerical simulations based on real data obtained from the current COVID-19 pandemic.
- Dr. Carlo Cenedese (ccenedese@ethz.ch)
- Dr. Andrea Iannelli (iannelli@control.ee.ethz.ch)
- Dr. Carlo Cenedese (ccenedese@ethz.ch) - Dr. Andrea Iannelli (iannelli@control.ee.ethz.ch)