Pre-Conference Workshops

The 63rd IEEE Conference on Decision and Control (CDC 2024) will be preceded by workshops on Sunday, December 15, 2024, addressing current and future topics in control systems presented by experts from academia, research institutes, and industry. The workshops will take place at the same venue.

1: Contraction Theory for Systems, Control, Optimization, and Learning

Abstract: Recent research has increasingly focused on applying the Banach contraction principle in the broad area of systems and control. Similarly, this tool plays a key role in addressing timely problems in machine learning and dynamical neuroscience. Contracting dynamical systems inherently offer numerous safety and stability guarantees. Additionally, the theory of monotone operators in optimization theory serves as an important complement to these theoretical tools.

The workshop will feature an extensive list of presentations by leading scientists from around the world on:

• The foundations of contraction theory
• Theoretical developments for complex networks, including advances in synchronization and scalability
• Computational advances in the design of contraction metrics and contracting dynamical systems for solving optimization problems
• Applications to machine learning, planning, and robust control

Of particular interest to the CDC audience will be findings on robust stability analysis and control design for both deterministic and stochastic systems, as well as formal robustness and stability guarantees for various learning-based control problems.

This workshop will bring together experts from diverse backgrounds to discuss recent theoretical and computational advances, identify emerging challenges, and explore rapidly-developing application opportunities. It should appeal to both junior and senior researchers interested in systems, control, and learning. The control community's interest in these topics is evidenced by recent well-attended events, including a tutorial session at the 2021 IEEE CDC and a pre-conference workshop at the 2023 ACC.

2: Control and Optimization in the Probability Space

Abstract: Driven by the recent advances and surge of interest in computational optimal transport and distributionally robust optimization within the machine learning and operations research communities, this workshop seeks to gather leading experts whose work lies at the crossroads of control theory and these emerging disciplines. This intersection of disciplines holds the promise to revolutionize the way we design high-performance control systems able to handle uncertainty in nonlinear, non-stationary and stochastic environments. The intersection between control, optimal transport and the distributionally robust paradigm indeed offers a fertile ground of new exciting theoretical challenges and modern real-world applications.

In this context, this full-day workshop will enable the interaction between researchers in the areas of stochastic control, computational optimal transport, and distributionally robust optimization, with the aim of: (i) developing a deeper understanding of the fundamental ties between these related research topics; (ii) leveraging this understanding to design optimal control algorithms able to handle uncertainty in nonlinear, non-stationary and stochastic environments. In doing so, the workshop will focus around two research directions:

• Control and Optimization in the face of distributional uncertainty. Classical approaches in control and optimization assume that uncertainty can be modeled either in a robust (deterministic, norm-bounded) or a stochastic (one fixed distribution, e.g., Gaussian) fashion. This assumption constitutes a big limitation in modern data science applications, where often only a finite amount of samples from the uncertainty is available. In such situations, one is confronted with distributional uncertainty, whereby not only is the system affected by uncertainty, but also the underlying probability distribution of the uncertainty is unknown and only partially observable. This issue led to the development of a novel class of distributionally robust uncertainty models, termed ambiguity sets, which can be constructed in a principled fashion from partial statistical information about the uncertainty (e.g., samples). This first research direction aims at addressing the new challenges introduced by ambiguity sets, namely the computational tractability, statistical soundness, and closed-loop guarantees of distributionally robust optimization and control problems.
• Dynamics and control in the space of densities. Stochastic dynamical systems are naturally represented in the space of densities via the continuity equation or, more generally, via the Fokker-Planck equation. This viewpoint provides several benefits, presenting a macroscopic outlook on the system by focusing on the ensemble behavior rather than individual trajectories. As an example, in large multi-agent systems, like those modeling social networks or cellular dynamics, the density representation facilitates the description and analysis of the entire population's behavior, contrasting with the original state-space model which typically addresses only individual agents/cells. However, transitioning from the standard Euclidean space to the space of densities presents numerous challenges, ranging from fundamental issues like controllability and stability to practical computational considerations. This second research direction will dive into these challenges through the lens of optimal transport, which has recently emerged as a powerful language to pose and address these problems.

Throughout the workshop, the methodological concepts will be motivated and illustrated by a myriad of exciting applications in machine learning, energy systems, network systems, and computational medicine. As apparent from the list of talks and the proposed schedule, the workshop has been built to be highly interactive and suitable for an audience with a diverse mix of academic/industrial backgrounds. Ultimately, our aim is indeed to create a vibrant space that appeals to both academic scholars and industry professionals, fostering a rich exchange of theoretical insights and practical applications.

The schedule for this full-day event is organized around 4 interlinked tracks (see below). The workshop will consist of talks of 25 minutes plus 5 minutes for Q&A after each individual talk. Besides this, we also built in the schedule breakout sessions to promote interactions. These breakout sessions will be a space for students to give spotlight presentations (3-5 minutes) where they could showcase their late breaking results (if you are a student interested in giving a spotlight presentation please contact one of the organizers).

4: Safe and Secure Learning-Enabled Systems

Abstract: Harnessing the power of machine learning to continuously monitor and detect anomalies advances the state of the art in instrumentation control. Learning-enabled systems have been rapidly increasing in size and acquiring new capabilities. These systems are typically deployed in complex operating environments, so their safety becomes extremely important. Ensuring safety requires that systems are robust to extreme events while we can monitor them for anomalous and unsafe behavior. While traditional machine learning systems are evaluated pointwise with respect to a fixed test set, such static coverage provides only limited assurance when exposed to unprecedented conditions in complex operating environments. One key question that remains unanswered is "how can we design and deploy learning-enabled systems that can be robust to extreme events while monitoring them for anomalous and unsafe behavior?" Indeed, given the increasing deployment of learning-enabled systems in various critical applications, guaranteeing security and safety of these systems has been an active research topic in different communities (e.g., control and machine learning) and has received a great interest from many funding agencies (e.g., NSF through the recent Safe Learning-Enabled Systems program).

The objective of this workshop is to bring leading researchers (including 2 NAE members and 3 female professors) in safe and secure learning-based control and verification, to discuss the latest developments, future directions, and explore possible novel directions in the intersection of learning, optimization, and game theory areas.

The workshop will feature one plenary talk (1 hour) and 9 half an hour talks. The workshop will conclude with a panel discussion on future research topics for safety and security of learning enabled systems.

We believe that with this line of diverse speakers our workshop not only will provide fruitful discussions on multiple angles of security and safety for learning-enabled systems but also will encourage the high attendance of various researchers/students from different groups, especially female participants.

Schedule

5: Control Architecture Theory

Abstract: The design and control of complex systems stands out as one of the paramount challenges of this century. Such systems are labeled as complex not only due to the intricacies of their individual components, but also because their functioning hinges on complex interactions among these components, across domains and scales.

To give a sense of the kind of systems we are interested in, think about the complex circuit governing a sensor employed in autonomous driving context, and autonomous vehicle which leverages the sensor, as well as a number of other complex hardware and software components within the autonomy stack, a fleet of autonomous vehicles of this kind, deployed and controlled following certain objectives, and interacting via a complex patterns, and a mobility system leveraging Autonomous Mobility-on-Demand (i.e., the fleet) systems as well as standard transit options.

Each of these subsystems is complex to design and control per se, and is influenced and influences other ones at different scales.

In this workshop we are driven by the need for a robust theory concerning layered control architectures (LCAs) across various complex systems, ranging from power systems and communication networks to autonomous robotics, bacteria, and human sensorimotor control.

Such systems exhibit exceptional capabilities, yet lack a cohesive, compositional theory for analysis and design, primarily due to their diverse domains. Conversely, there exists a fundamental set of control concepts and theories which are universally applicable and can accommodate domain-specific adaptations. In this context, however, control methods are often limited to work for individual components of larger systems, lacking comprehensive theoretical foundations.

Although fragments of a control architecture theory have emerged across disparate disciplines and domains, a unified theory and community are lacking. Against this backdrop, the objective of this workshop, organized as a companion to an invited session on the same subject, is to cultivate a new interdisciplinary community which considers control architectures and systems theory as a central focus of study. In particular, we will bring together experts from a wide range of disciplines, spanning robotics, control, applied mathematics, systems biology, aerospace, etc., with the goal of establishing a common language and core set of challenge problems and techniques.

A key outcome of the workshop will be a white paper laying out a research program in this area.

Schedule: 8:45 AM - 5:30 PM, and speakers/panelists include Aaron Ames (Caltech), Domitilla Del Vecchio (MIT), John Doyle (Caltech), Florian Dörfler (ETHZ), Nadia Figueroa (UPenn), Nikolai Matni (UPenn), Lisa Li (Michigan), Manfred Morari (UPenn), Alberto Sangiovanni-Vincentelli (UC Berkeley), Alberto Speranzon (Lockheed Martin), Paulo Tabuada (UCLA), and Gioele Zardini (MIT).

The correct order of the talks (30 minutes each) is being determined, and will appear on the webpage. The workshop will end with a panel discussion.

6: Leveraging bifurcations for control of intelligent and collective behaviors

Abstract: From single-celled organisms to animal groups and human societies, living beings across scales sense their environment and react to it adaptively for survival. In doing so, they provide compelling examples of control systems capable of generating extremely robust and yet adaptable intelligent behaviors. Growing evidence has pointed to nonlinearity and, in particular, to bifurcations in nonlinear models as key ingredients for understanding the robust adaptability of these systems. To engineer autonomous systems that provably inherit the robust adaptability of their biological counterparts, control engineers must leverage these scientific insights and embrace bifurcation as a design principle.

Bifurcations have a constructive role in natural and societal intelligence. A local bifurcation point is a parameter regime at which a solution of a nonlinear system changes stability. This is a point of ultra-sensitivity at which a control system can rapidly change its behavior in response to changes in the environment, even when those changes are arbitrarily small. This ultra-sensitivity can be observed in human decision-making, for example when a sudden event that requires a change in behavior happens while riding a bike or cooking a meal. Collective behaviors also exhibit ultra-sensitive responses, e.g., as a bacterial society does when antibiotics are poured into its environment, or as a human society can choose to do when imminent dangers are around the corner. Away from the ultra-sensitive bifurcation point, a control system ruled by bifurcations is organized into distinctively different robust behaviors associated with choosing a control action or strategy over alternative ones: steering the bike left instead of right into a pedestrian; turning the stove off instead of burning the sauce; developing a biofilm instead of remaining exposed to antibiotics; transitioning into a more sustainable life style instead of doing business as usual. The co-existence of many different possible control choices is captured by the rich multi-stable attractor landscape that emerges at bifurcations. By navigating this landscape in response to inputs or in pursuit of goals, an agent can continuously adapt its behavior in a robust yet sensitive fashion. In other words, bifurcations can be leveraged for control to achieve robust and adaptive behaviors in ever-changing and unpredictable environments. This is diametrically opposed to the classical approach of controlled bifurcations for stabilization problems, in which a typical objective is to steer a system away from a bifurcation point, ideally achieving global stability.

To understand natural and societal collective and intelligent behaviors, and to draw inspiration from them for the design, analysis, and control of more robust and adaptable artificial intelligent and collective behaviors, a new synthesis of bifurcation and systems theories is needed. This workshop aims at providing the state-of-the art of recent efforts toward this new synthesis, and to explore new ideas toward its realization in applications.

7: Control of Multiagent Systems: Challenges and Solutions

Abstract: Over the last two decades, technology has significantly advanced the development of integrated systems that combine mobility, computing, and communication on a single platform. As a result, we have rapidly entered a new era in which teams of agents, known as multiagent systems, interact with each other to influence their motions for cooperatively performing a wide array of civilian and military applications. These applications range from surveillance and reconnaissance to unmanned system operations and energy management. It is therefore not surprising that the overall robotics market value has seen a dramatic increase with the expectation to reach around 200 billion U.S. dollars by 2025. This has led to a significant research activity focused on how to control these robot teams through local interactions (i.e., distributed control) to achieve necessary cooperative behaviors.

Specifically, distributed control approaches can typically be classified into two categories; namely, "leaderless distributed control approaches," where all agents perform a task without an external command, and "leader-follower distributed control approaches," where a subset of agents (referred to as leaders) receive external commands that influence the behavior of other agents (referred to as followers) in the multiagent system. The objective of this workshop is to cover the state-of-the-art advancements in the field of multiagent systems with a focus on leaderless and leader-follower distributed control approaches. Participants will have the opportunity to learn about the challenges and solutions on problems related to a) spatiotemporal, communication, and nonholonomic constraints; b) multiagent system resilience against uncertainties and time-delays; and c) autonomy including coordination, collaboration, optimization, and decision-making with applications to mobile ground, aerial, and space robots as well as energy systems.

Constraints. Spatiotemporal constraints refer to the limitations on the movement and actions of agents that are imposed by both space and time. These constraints can include factors such as physical boundaries and timing requirements for coordinated tasks. Furthermore, communication constraints refer to the limitations on the exchange of information between agents, where these constraints can arise due to limited communication range, bandwidth, or signal interference. Nonholonomic constraints are also restrictions on the motion of agents, where the direction of motion is limited by the steering mechanism. Addressing all these constraints is crucial for effectively using multiagent systems in real-world applications.

Resilience. Achieving resilience in real-world applications is also highly important for the safe operation of multiagent systems. In this context, uncertainties and time-delays play a crucial role in threatening safety as they can lead to unpredictable behavior and delayed responses. In particular, uncertainties can arise from a broad spectrum of sources including unpredictable environmental conditions and modeling inaccuracies, while time-delays can occur in agent-to-agent communication and control loops. Ensuring resilience against these factors is key to maintaining multiagent system stability and predictable performance.

Autonomy. Autonomy requires coordination, collaboration, optimization, and decision- making with no or minimal human intervention. Specifically, coordination and collaboration are central to the autonomous operation of multiagent systems. Typically, the strategy involves dividing team-level tasks into manageable subtasks with each agent responsible for executing its assigned portion in a coordinated manner. Yet, by integrating agents with diverse capabilities, the potential arises to unlock entirely new functionalities and skills. Moreover, optimization for autonomous motion planning and resource allocation is at the heart of effective decision- making. By properly formulating and solving optimization problems, agents have the ability to determine the most efficient ways for achieving their objectives while adhering to constraints and considering the actions of other agents.

Covering the topics on constraints, resilience, and autonomy is crucial for the advancement and success of the next-generation of multiagent systems, where we organize this workshop on these key topics. As given in the schedule below, Talk 1 is related to spatiotemporal constraints, Talks 2 and 3 are related to communication constraints, and Talk 3 is also related to nonholonomic constraints (i.e., these talks are related to theme a) of this workshop). In addition, Talks 4 and 5 are related to multiagent system resilience against uncertainties and time-delays (i.e., these talks are related to theme b) of this workshop). Furthermore, Talk 6 is related to coordination and collaboration, whereas Talks 7-11 are related to optimization and decision-making aspects in autonomy (i.e., these talks are related to theme c) of this workshop). Finally, it is important to note that Talks 1, 2, 6-11 will further touch upon multiagent system applications to mobile ground, aerial, and space robots as well as energy systems.

8: Fair Decision-Making and Societally-Aware Control in Networked Systems

Abstract: This workshop aims to bring together researchers and scientists to discuss topics related to the theory of control of networked systems, addressing pressing societal challenges such as promoting fairness in control strategies and accounting for the impact of feedback and social engagement in interaction-based decision-making. The workshop is designed with a strong interdisciplinary component, and the talks, organized in four separate sessions, include relevant topics that combine classical control techniques with game theory and opinion dynamics modeling, emphasizing the practical application of theoretical frameworks to real problems. A round table at the conclusion of the day will offer significant insights into the latest developments in the field, open research directions, and potential applications in mobility, economics, and recommendation systems.

9: Large Population Teams: Control, Equilibria, and Learning

Abstract: Stochastic teams entail a collection of decision makers / players /agents acting together to optimize a common cost function, but not necessarily having access to the same information. It is by now known that such problems are challenging owing to information structure-dependent subtleties even in the team setup. In the game setup, informational dependence makes solutions very fragile in that even solution concepts need to be refined, the value of information can be negative (unlike in collaborative teams), the value of information is not necessarily continuous, and the presence or absence of common randomness turns out to be a key attribute. Large-population games become even more challenging when one or more competing teams are involved. There are two major challenges when trying to solve such competitive team problems: First, large-population team problems are computationally challenging since the solution complexity increases exponentially with the number of agents, and, in general, the team optimal control problems belong to the NEXP complexity class. Second, competitive team problems are conceptually challenging due to the elusive nature of the opponent team, and thus one cannot directly deploy approximation techniques available in the large-population game literature. Despite the great advances in mean-field approximations for multi-agent systems for specific classes of single team games, several unresolved challenges still exist for the more realistic case of competitive team problems.

The proposed workshop intends to address the above needs and by bringing ogether researchers from various disciplines (different areas of engineering, mathematics, and data science) working on the theory and applications of decentralized systems with large number of units or agents under a variety of system and evolution dynamics, information structures, performance criteria, and application areas. A common thread will be to understand the optimality and equilibrium behaviour, scaling behaviour with the number of agents, and learning dynamics; all in both the associated mathematical theory as well as in the context of emerging engineering and applied science applications. Due to the interdisciplinary nature of such problems involving optimal control, stochastic control, game theory, multi-agent systems, and robotics, we intend to establish connections between various formulations adopted in the community and bring researchers who have been using alternative setups, solutions approaches, and applications to allow for exchange of ideas and formulation of new research directions and collaborations. One other main goal of this workshop is to inspire a future generation of researchers in this vibrant field.

10: Complex Socio-Technical Networked Dynamics

Abstract: Complex socio-technical systems are increasingly becoming a cornerstone of our societies, playing a key role in shaping numerous facets of our daily life, from production and communication, to robotics and information accessibility. A key feature of such systems in the presence of intertwined layers involving technology, human behavior, and infrastructure networks. The pervasive presence and impact of these systems in our societies calls for a deeper understanding of such complexity towards the development of a unified platform to explore the multifaceted dimensions of socio-technical systems.

Motivated by such a key challenge, we propose a workshop with the primary objective of creating a platform for researchers to discuss innovative ideas in the realm of complex socio-technical systems. To this aim, we selected a list of speakers from research units specialized in diverse aspects of modeling, analysis, and control of complex socio-technical systems, with a broad spectrum of expertise spanning from the development and analysis of mathematical models and control protocols for cyber-physical-human-systems, complex social networks, and the exploration of human-robot interactions. The list of speakers will be complemented by a conclusive session in which junior researchers (PhD students and postdocs, which will be selected via an open call) will have the opportunity to illustrate and discuss their innovative approaches to complex socio-technical networked systems.

This workshop will offer a diverse yet consistent and thorough overview of the ongoing efforts towards hacking complex socio-technical systems, showcasing how systems- and control-theoretic approaches offer powerful tools to understand, model, and control such dynamics, while bringing innovation from other fields, such as evolutionary game theory, network science, experimental psychology, and robotics. vibrant field.

11: Past and Future in Control of Networked Systems: Insights and Perspectives

Abstract: The objective of this workshop is to provide insight and perspectives on the area of control of networked systems. The workshop coincides with the 10-year anniversary of the IEEE Transactions on Control of Networked Systems. Over the past fifteen years or so important strides have been made in research on decision and control systems characterized by a distributed or networked architecture, be it in modeling, analysis, estimation, design, or implementation. New research avenues have been opened and established: from collaborative control, distributed learning, multi-agent systems, distributed optimization, control of collective behavior, distributed estimation, game theory, dynamical systems over graphs, coevolutionary networks, synchronization, large-scale complex systems, to control with communication constraints. Application areas relevant to control of network systems include smart infrastructure, multi-robot systems and swarm robotics, systems biology, neuroscience, smart health, computing, communications, transportation, manufacturing, power systems, cyber-physical and social systems, sensor networks, and social networks.

The workshop aims to provide the framework for an introspective about the past and future of network control. Featuring invited talks from leading researchers, the workshop be a combination of recent advances, with a tutorial flavor (e.g., distributed optimization, distributed control, game theory) and application areas (e.g., cyber-physical systems, energy, transportation, swarm robotics, social networks), with a forward-looking outlook. We have assembled a set of speakers whose exciting research in these areas will shed novel insights. Our deliberate selection of speakers aims to facilitate the cross-fertilization of ideas, fostering an environment conductive to the identification of open challenges in the area of network control.

12: From Formal Methods to Data-Driven Verification and Control

Abstract: Formal verification and controller synthesis for dynamical systems have garnered remarkable attention over the past two decades, driven by their extensive applications in safety-critical systems. While these formal approaches have become indispensable across numerous applications, they often necessitate closed-form mathematical models of dynamical systems. However, these models might either be unavailable or too complex to be constructed in real-world scenarios. Hence, the use of data-driven techniques becomes essential in enabling formal analysis for systems with unknown dynamics.

Over the past decade, several data-driven techniques have been proposed for the formal verification and controller synthesis of unknown dynamical systems. One may classify them in two types: the indirect and direct approaches. More specifically, indirect data-driven techniques are those which leverage system identification to learn approximate models of unknown systems, followed by model-based controller analysis approaches. In comparison, direct data-driven techniques are those that bypass the system identification phase and directly employ system measurements for the verification and controller design of unknown systems.

In this workshop, we bring together a number of researchers active in the area of data-driven verification and control with provable guarantees. Along with cherishing the exchange of ideas between researchers in the field, by gathering a number of key talks we aim to achieve the following goals for the audience attending the workshop:

• Provide to the researchers new to these topics an updated view of the state of the art on data-driven verification and control including indirect and direct data-driven techniques.
• Discuss scalable directions and areas to mitigate the, so-called, sample complexity.
• Suggest novel applications of these data-driven techniques.

13: Data-driven modelling, analysis, and control using the Koopman operator

Abstract: The linearity of Koopman operators and the forecast simplicity of linear time-invariant (LTI) models coming from their functional representation in a space of "observables" lead to their increased popularity for learning dynamical systems. This representational simplicity inspired a bevy of system identification approaches and holds promise for solving many classes of nonlinear control problems through lifting nonlinear systems into suitable spaces of observables. Given that operator-theoretic system identification methods are still under active development, it is critical to present current results, main challenges and point to fruitful directions for making Koopman-based approaches more mature for systems-and-control applications.

This workshop aims to present a broad overview of state-of-the-art Koopman operator approaches from the perspective of systems and control theory. The main objective is to provide a multifaceted perspective through an introduction to fundamental concepts, current opportunities/challenges, and recent advancements – striking a balance between a tutorial style and presenting the latest advancements.

The workshop is organized around the following four thematic areas:

• A tutorial introduction, (im)possibilities and dynamic mode decomposition (DMD)
• Choosing observables: from exact embeddings, and kernel methods to representation learning
• Perspectives on optimal, predictive, and robust control
• Certifiable learning and control design

14: Data-driven control: theory and applications

Abstract: Data-driven and predictive approaches are increasingly popular in control theory and its applications. Among the reasons for the predominance of data-driven perspectives in current research are the large amounts of data generated by to-be-controlled plants, the complexity of the system dynamics and the available large computational power. Each of these motivations brings with it specific challenges, for example developing efficient algorithms and dealing with uncertainties and inaccuracies.

The objective of the workshop is to provide a review of some of the origins, a critical evaluation of some aspects of the state-of-the-art, and some perspectives on current practical applications. To this purpose we plan to assemble experts in the theory and practice of data-driven and predictive control methods to present their experiences and point of views.

The workshop is organized around the following four thematic areas:

• Behaviors and data
• Predictive control
• Uncertainty and robustness
• Data-driven control applications

The workshop consists of 12 talks by experts in the field. There are three talks associated with each of the themes.

15: Learning Dynamics from Data: Fusing Machine Learning and System Identification

Abstract: Recent decades have witnessed significant advancements in the field of system identification. By analyzing input-output data and extracting mathematical models, system identification facilitates the characterization and prediction of dynamics, aiding in the subsequent controller design, optimization, and decision-making procedures. In traditional system identification, the focus is on learning a model of a specific system through pre-existing physical knowledge and measured input-output trajectories. This typical workflow is inextricably linked to supervised machine learning, i.e., discovering mathematical relationships in the data. Hence, we recognize that fusing system identification and machine learning presents a remarkable opportunity. This fusion enables traditional system identification to efficiently handle nonlinear systems and extend machine learning to be applicable to commonly encountered complex dynamical systems in real life. However, the challenges revolving around fusing machine learning with system identification, as evidenced by current results, primarily center on two aspects: i) lacking physical interpretation; ii) leveraging knowledge accumulated across related systems and tasks. We also focus on the opportunities that this opens in terms of "new" problems that have appeared: transfer learning, learning dynamics during reinforcement learning and model predictive control, etc.

The workshop offers a repertoire of the most recent theoretical and practical developments of learning dynamical systems from data, both from the more classical system identification point of view as well as from a machine learning perspective. We also aim to cover a wide diversity of application domains, ranging from classical systems and control, to robotics, to general scientific learning. The key topics include: i) physics-informed learning; ii) meta-learning, and iii) deep neural networks, with applications in i) building energy systems; ii) electromechanical systems, iii) mechatronic systems, etc. Beyond the presentations, ample time will be provided for discussion together with the audience. Furthermore, the workshop seeks to foster valuable interdisciplinary dialogue and collaboration between academia and industry.

16: Remembrance of Allen Tannenbaum: Foundations of modern robust control and beyond

Abstract: On December 28, 2023, a remarkably creative academic career came to an end. Allen R. Tannenbaum left behind an astonishingly vast corpus that spans many fields, including feedback theory, optimal and robust control, signal processing, image processing and biomedical imaging, robotics, operator theory, algebraic geometry and other topics in abstract mathematics, the theory of optimal mass transport, network science, and many more. The present workshop aims to sample from this vast and diverse opus and broader related fields, focusing in particular on the field for control, to celebrate Allen's legacy and inspire the coming generation of control theorists and scientists.

Allen began his academic life in algebraic geometry, trained in pure mathematics at Harvard under Heisuke Hironaka. Over a period of almost half a century he went on to create new fields and impact science, engineering, biology and many other fields in a multitude of ways. In the late 1970's he introduced and solved the problem to optimize gain margin for linear systems using analytic interpolation theory. This foundational contribution to the field of modern robust control was a harbinger of a great chapter that began to emerge and kept the community captive throughout the 1980's, culminating in what is now known as H-infinity-control. His book on Feedback Control Theory, co-authored with John Doyle and the late Bruce Francis, nicknamed DFT, has been a standard reference for the last three decades.

Allen pioneered the use of partial differential equations in computer vision and biomedical imaging, and the applications of Optimal Mass Transport theory to image analysis, network science, systems biology, and cancer genomic analysis. A long list of seminal contributions in the aforementioned topics and the broader field of systems and control led to transformative insights that will guide and inspire generations to come.

The purpose of the workshop is to celebrate the life and legacy of Allen. Those who knew personally will miss him dearly. The broader community will continue to be inspired by his work and his example. The workshop will consist of a series of talks on technical subjects by Allen's colleagues, co-workers, students and professional friends, interspersed with historical remarks and anecdotes on the ideas, influence, and science of Allen.

17: Control and Adaptation: Imagine What’s Next - Celebrating the 60th Birthday of Miroslav Krstic

Abstract: The workshop "Control and Adaptation: Imagine What’s Next" reviews the forefront of research on topics ranging from nonlinear control and partial differential equation (PDE) control to extremum seeking control (ESC) and games. The agenda is co-headlined by two of the most eminent scholars on these subjects: Professors Tamer Başar and Jean-Michel Coron. The first segment of the workshop covers control design methodologies for PDEs: backstepping, Lyapunov-based methods, designs for PDE-ODE interconnections, adaptive control, and flow control in various domains such as traffic, phase change in materials, and oil drilling processes. The second segment covers advances in extremum seeking and control of delay systems. Besides the exposure for students to the state of the art and historical evolution of key research subjects, the workshop will foster an atmosphere conducive to the formation of new partnerships and collaborations.

The workshop is held on the occasion of Professor Miroslav Krstic’s 60th birthday. It celebrates his pioneering role in the creation of these research subjects and his legacy of nurturing the respective communities of researchers.

18: NedicFest! A Workshop Celebrating Angelia Nedić at the CDC 2024

Abstract: This workshop is being organized to celebrate Professor Angelia Nedic’s 60th birthday and honor her multiple long-lasting contributions to the field of optimization and control theory. This workshop brings together 21 of her colleagues, collaborators, and former students and postdocs (including organizers and invited speakers) who will present a broad range of contemporary topics in different areas of optimization and control theory. The main goal of this workshop is to inspire a future generation of research leaders to pursue work that promotes excellence and will thus likely have a profound impact in the field.