For participants > Tutorials

The tutorials will take place in the afternoon of Monday July 18th.

Three tutorials will be offered:

1. Multiscale Modeling and Control of Infectious Diseases
2. UNICORN - A Unified Control Framework for Real-Time Power System Operation
3. Mean-Field-Type Games for Engineers

 ________________________________________________________________________________________________________________________________________________

Tutorial: Multiscale Modeling and Control of Infectious Diseases

Organizers:

Prof. Esteban A. Hernandez‐Vargas
Institute of Mathematics, UNAM, Mexico 
Frankfurt Institute for Advanced Studies, Germany 

Prof. Alejandro H. González
Argentine National Scientific and Technical Research Council ‐ INTEC (CONICET‐UNL), Argentine 

Prof. Antonio Ferramosca 
University of Bergamo, Italy

Summary:

Biomedical engineering is a multidisciplinary field linking the application of engineering principles and  tools to medicine and biology for healthcare purposes. Furthermore, pharmaceutical companies have taken a strategic initiative to promote the use of modeling approaches within drug projects. The value of a model‐ based approach to drug development for improved efficiency and decision-making at preclinical development phases has been largely advocated. Drug administration is classically divided  into two phases, a so-called pharmacokinetic (PK) phase that relates dose, frequency, and route of  administration to drug level‐time relationships in the body, and a pharmacodynamics (PD) phase that  relates the concentration  of the drug at the sites of action of the magnitude of the effects produced. 

This workshop aims to provide the basic principles of control, modeling, and biology to: 

• understand biomedical engineering
• provide a vision of the applicability to infectious diseases
• develop mathematical models and control strategies in medical applications

These contributions are meaningful because they allow building a systematic framework that can be applied to simulate, emulate, and control other diseases. Furthermore, recent modeling advances will be presented in different viral infections dissecting detailed contributions of key players to severe viral infections as well as their respective interactions are crucial for developing treatment strategies.  In the same manner, advances in fully‐automated systems for insulin release are discussed in detail; as well as current challenges in COVID‐19. 

The results that will appear in the workshop include simulations as well as rigorous mathematical analyses that guarantee control engineering strategies properties. Ultimately, simulations of PK/PD are introduced and discussed to evaluate dose‐concentration‐response and predict the effect‐time courses resulting from the treatment.

Content:
Introduction to biomedical systems and control
Mathematical Modeling and Parameter fitting with differential evolution algorithm
Optimal and suboptimal control strategies for infectious diseases at the host level
Optimal control for SIR-type systems
Concluding Remarks

The intended audience of this workshop consists of researchers, control practitioners, MSc and Ph.D. candidates as well as any scientist who wants to get in touch with this field. 

Speakers:

Prof. Esteban A. Hernandez‐Vargas
Institute of Mathematics, UNAM, Mexico 
Frankfurt Institute for Advanced Studies, Germany 

Prof. Alejandro H. González
Argentine National Scientific and Technical Research Council ‐ INTEC (CONICET‐UNL), Argentine 

Prof. Antonio Ferramosca 
University of Bergamo, Italy 

 ________________________________________________________________________________________________________________________________________________

Tutorial: UNICORN - A Unified Control Framework for Real-Time Power System Operation

Organizers:

M. Sc. Lukas Ortmann, ETH Zurich

Dr. Saverio Bolognani, ETH Zurich

Dr. Jean Maeght, RTE France

Summary:

Due to unprecedented changes in power systems, new real-time approaches that enable optimal operation are essential. Feedback Optimization is such a method that turns optimization algorithms into feedback controllers. Utilizing the advantages of both optimization and control provides us with a real-time compatible controller which make (power) systems track the optimum of an optimization problem, while satisfying constraints. We present the theoretical basis and the results of a three-year project between ETH Zurich and the French transmission grid operator RTE. The tutorial aims at theoretical researchers interested in new research directions in the field of Feedback Optimization as well as practitioners who would like to learn how to apply Feedback Optimization in power systems. 

Content:

Introduction to the theory of Feedback Optimization (Saverio Bolognani)
Experimental validation of Feedback Optimization for power systems including the Unicorn 4-bus benchmark (Lukas Ortmann)
Statistical analysis of Feedback Optimization vs state-of-the-art control including the Unicorn 56-bus benchmark (Saverio Bolognani)
Transmission Grid Operation with Feedback Optimization including the Unicorn 7019-bus benchmark (Lukas Ortmann & Jean Maeght)
Panel Discussion and Q&A

The Unicorn project is a three-year research collaboration between the French transmission grid operator RTE and ETH Zurich funded by the Swiss Federal Office of Energy. More details can be found at www.unicorn.control.ee.ethz.ch. There, we will also publish the code for the benchmarks.

Speakers:

M. Sc. Lukas Ortmann, ETH Zurich

Dr. Saverio Bolognani, ETH Zurich

Dr. Jean Maeght, RTE France

 ________________________________________________________________________________________________________________________________________________

Tutorial: Mean-Field-Type Games for Engineers

Organizers:

Julian Barreiro-Gomez
NYUAD Research Institute, New York University Abu Dhabi
Center on Stability, Instability and Turbulence (SITE)

Hamidou Tembine
NYUAD Research Institute, New York University Abu Dhabi
Center on Stability, Instability and Turbulence (SITE)

Summary:

A class of stochastic differential games, known as mean-field-type games, incorporates the distribution of the variables of interest, e.g., the strategies and/or the system state, into the analysis. This feature allows us to take into consideration risk terms such as the variance or even higher-order terms. In general, the solution for this kind of stochastic game problems is complex and requires solving a backward partial integro-differential equation (PIDE), corresponding to the Hamilton-Jacobi-Bellman equation, coupled with a forward partial differential equation describing the evolution of the distribution of the states, known as the Fokker-Plank-Kolmogorov equation. In this tutorial session, we avoid addressing such complex PIDE system and solve the underlying stochastic differential game in a semi-explicit way by proposing an appropriate ansatz for the value function and following the so-called verification/direct method. This tutorial session is mainly designed for new researchers on the topic and beginners in the area. The topic will be addressed in a friendly way trying to present the contents in a quite easy manner. 

Content:

Introduction to Mean-Field-Type Games
Semi-Explicit Solution for LQ Mean-Field-Type Games: Non-cooperative, Cooperative and the Coopetitive Solutions
Semi-Explicit Solution for Some Non-Linear Mean-Field-Type Games
Engineering Applications

Speakers:

Julian Barreiro-Gomez
NYUAD Research Institute, New York University Abu Dhabi
Center on Stability, Instability and Turbulence (SITE)

Hamidou Tembine
NYUAD Research Institute, New York University Abu Dhabi
Center on Stability, Instability and Turbulence (SITE)

________________________________________________________________________________________________________________________________________________