ASPIRE CPDS Workshop on
Control and Optimization of Complex Systems

Jérémie Kreiss received a double M.Sc. and Engineering degree in Electrical Engineering from Université Claude Bernard Lyon 1 and INSA Lyon in 2016. He completed his Ph.D. in Automatic Control at INSA Lyon in 2019, receiving the 2020 Best Ph.D. Award for his work on Energy for Sustainable Development. After a year as a Teaching Assistant at École Centrale de Lyon, he joined Université de Lorraine as an Assistant Professor in 2020. A recipient of the 2025 ANR JCJC grant, he serves as an Associate Editor for the ICSTCC and on the IEEE CSS Conference Editorial Board. His research focuses on geometric control theory and input-redundant systems applied to power electronics.

Pietro Lorenzetti is a Chargé de Recherche at CNRS, affiliated with CRAN, Université de Lorraine, France. He obtained his PhD in Automatic Control in  2023 from Tel Aviv University within the Marie Skłodowska-Curie Horizon 2020 ConFlex program. His research interests include the analysis and control of nonlinear  systems, with particular emphasis on integral control, singular perturbations, projected dynamical systems, and applications to power systems and power electronics.  He has been a visiting researcher at the University of Melbourne (Australia), Tampere University (Finland), TU Eindhoven (The Netherlands), and the Illinois Institute of  Technology (USA). He is a member of the IFAC Technical Committee on Nonlinear Control Systems and of the IEEE Control Systems Society Conference Editorial Board.

Ehsan Nekouei joined the Department of Electrical Engineering of City University of Hong Kong in 2019 as an Assistant Professor, where he was promoted to an Associate Professor in March 2025. Prior to that, he was a postdoctoral researcher at the KTH Royal Institute of Technology. His research interests include privacy in networked control systems and secure industrial control systems.

Vaibhav Upadhyay is a dual degree student at IIT Bombay, pursuing a B.Tech. in Aerospace Engineering and an M.Tech. at the Centre for Systems and Control. His research interests are in numerical methods for constrained feedback synthesis of nonlinear control systems and control of PDEs.

Lorenzo Zino is an Assistant Professor with the Department of Electronics and Telecommunications, Politecnico di Torino, Italy, since 2022. He received the BS, MS, and PhD in Applied Mathematics from Politecnico di Torino. He held Research Fellowships at Politecnico di Torino (Italy), University of Groningen (The Netherlands), and New York University (US), and visiting positions at Lund University (Sweden), Curtin University (Australia), and Adelaide University (Australia). His research interests include modeling, analysis, and control of dynamics over networks, applied probability, and game theory. He has co-authored more than 100 international scientific publications, including 60 journal papers. He is Senior Member of the IEEE and the recipient of the 2024 Best Young Author Journal Paper Award from the IEEE CSS Italy Chapter. He is member of the Editorial Board of Scientific Reports, Associate Editor of the Journal of the International Journal of Control and IEEE Control Systems Letters, and member of the CEB for IEEE CSS and EUCA.

Samuele Zoboli received his M.Sc. in Automation Engineering from the University of Bologna (Italy) in 2019 and his Ph.D. in Control Theory from the University of Lyon 1 (France) in 2023. Following a postdoctoral fellowship at LAAS-CNRS in Toulouse (France), he joined the institute as a CNRS Researcher in 2025. His research focuses on the stabilization of discrete-time nonlinear systems, multi-agent systems, and the intersection of reinforcement learning and control-applied artificial intelligence.

The talk introduces a direct optimal control framework for constrained minmax density transportation for noisy linear parabolic PDEs. The problem is reformulated as a convex semi-infinite program via spatial discretization and finite parametrization of control and disturbance trajectories. Techniques from convex semi-infinite programming theory are employed to characterize exact solutions that guarantee constraint satisfaction in continuous time across an uncountable family of disturbance realizations. The effectiveness of the proposed approach is demonstrated via numerical experiments.

In this talk, we will study privacy-aware state estimation for a system where a private process drives the state. The state estimates are shared with an adversary who might attempt to infer the private process from the state estimates. We formulate the privacy-aware estimation as an optimization problem that aims to minimize a linear combination of the estimation error and leakage of private information captured by mutual information between the private process and the estimator output. We show that the optimal estimator design problem is a closed-loop control problem where the estimator controls the adversary's belief. We also derive a certain backward optimality principle for the optimal estimation policy. The numerical design of the estimator is challenging due to the high computational cost of mutual information. To solve this problem, we develop an efficient algorithm that combines the policy gradient approach with a variational formulation of the KL-divergence. We finally compare the performance of the proposed algorithm with that of the differential privacy approach using a building automation application. 

For Over-Actuated systems, classical control allocation breaks down when actuators have dynamics.  Recent geometric control allocation frameworks overcome this limitation,  but their algorithms lose efficiency and guarantees in large-scale systems. In this presentation, after a brief review of classical control allocation, we introduce the notion of almost invisible inputs,  which makes it possible to quantify and control the approximation error,  offering a scalable and rigorous way to extend geometric control allocation.

Complex social systems are characterized by the presence of a deep intertwining between opinion formation and decision-making processes. In the literature, these two processes have been typically studied as separate problems, establishing different dynamical models to represent them. In this talk, I will present a novel modeling paradigm that bridges this gap, proposing a unified co-evolutionary model for actions and opinions that builds on a game-theoretic approach. First, through the analysis of the model, I will shed light on different emergent behaviors that our model can predict, including consensus formation, polarization, and emergence of unpopular norms. Second, I incorporate a control action within the model with the goal of steering the population, initially at a consensus, to a different consensus state. Building on our theoretical results, we establish a methodological approach to design optimal interventions in social systems.

Classical contraction theory offers a robust framework for nonlinear system analysis, providing global guarantees that all trajectories converge to a unique equilibrium without requiring that equilibrium to be known a priori. However, this stringent requirement often limits its applicability to systems exhibiting complex behaviors, such as multi-stability. To bridge this gap, the concept of k-contraction generalizes distance-shrinking to the contraction of k-dimensional volumes between trajectories. This talk introduces the core principles of contraction analysis, evaluating its inherent strengths and limitations, before transitioning to the geometric interpretation and stability guarantees of k-contraction. To address the computational burden typically associated with matrix compounds, the multilinear algebra tools classically used to study k-contraction, I propose analysis and design techniques based on matrix inequalities. In the framework of Lur’e systems, these machineries can be transformed into tractable LMI conditions. The presentation concludes by applying these methods to extremum control under non-convex optimization objectives.

In this talk, I will present an overview of my current research activities, with particular emphasis on projected dynamical systems and their role in constrained  control problems. Projected dynamical systems provide a mathematical framework for describing the evolution of dynamical systems under constraints. Originally developed  in connection with variational inequalities in economics and optimization, they also offer useful tools for control design and analysis. Motivated by applications in power systems,  I will discuss how this framework can be used in the analysis and design of constrained integral controllers to address safety and performance requirements. Rather than  focusing on technical details, the goal of the talk is to highlight the main questions, methodologies, and perspectives that motivate this line of research.