Tutorial Sessions

Non-Rational Control

Abstract: This tutorial aims to expose researchers to non‑rational control: a unified framework for designing practical, scalable, and stabilizing controllers for infinite-horizon problems that are otherwise analytically and computationally intractable, especially when the optimal solution is infinite-dimensional. We will unpack its three pillars: infinite‑dimensional convex duality, fast optimization algorithms in the Fourier‑domain, and a rational‑approximation pipeline that delivers finite‑order controllers with provable near‑optimality. Along the way, we will present a cohesive treatment of a wide range of control objectives, including distributionally robust, risk-sensitive, and mixed-criteria formulations, demonstrating how they all fit within a common paradigm. By the end, participants will be equipped with both the theoretical foundations and practical algorithms needed to efficiently design high‑performance finite-dimensional controllers for the most demanding modern applications.

Introduction to Nonstochastic Online Control

Abstract: This lecture describes online nonstochastic control (ONC), a new framework which interpolates between the stochastic and worst-case regimes of optimal control. The key innovations are (1) to replace stochastic perturbations with adversarial ones, and (2) to adopt the regret metric from the theory of online decision making to quantify performance. Building from this formalism, we introduce a range of computationally efficient algorithms which attain optimal rates of regret across a large breadth of regimes: partial observability, unknown system dynamics, learning in marginally stable systems. These advances not only resolve open-questions in the statistical complexity of adaptive control but have further inspired neural architectures for sequence modeling.

Data Driven and Learning Enabled Control

Abstract: Data-driven control (DDC), that is the design of controllers directly from observed data, has attracted substantial attention in recent years due to its advantages over model-based control (MBC). DDC avoids a computationally expensive, potentially conservative model identification step and bypasses practically difficult questions such as model order/class selection. The objectives of this session are: (1) Expose the audience to a slate of recently developed methods, covering both analytic approaches and learning enabled ones, indicating the relative strengths of each. (2) Provide a key to the rapidly expanding literature in the subject, to help researchers newly interested in this field to quickly come up to speed. (3) Point out to open problems at the confluence of control and machine learning, and indicate how control-theoretic tools can help in analyzing the properties of newly proposed data-driven machine learning techniques. The proposed session will be structured around four 25 minutes talks, each covering a different aspect of Data Driven Control. The first two talks will cover analytical approaches, while the last two cover machine-learning based ones. The last 20 minutes of the session will be used for an open discussion on data driven control and the role of control theory on newly developed data-driven machine learning approaches such as structured state space (Mamba).

Sampling-Based Methods for Optimal Control: Theory, Algorithms, and Applications

Abstract: Optimal control problems for real-world nonlinear systems are frequently characterized by highly nonconvex or even discontinuous objective functions and dynamics, particularly in domains like contact-rich robotics. Recently, Sampling-Based Optimal Control (SBOC) methods have gained considerable popularity. Their appeal stems from the ability and flexibility to handle highly nonlinear or even discontinuous dynamics and complex cost functions effectively. Furthermore, their inherent structure allows for massive parallelization, especially on modern hardware like GPUs. SBOC has found successful applications in diverse areas, including path planning, robotic control (e.g., legged locomotion, manipulation), and model-based reinforcement learning. The primary objectives of this tutorial are to provide attendees with a comprehensive overview of Sampling-Based Optimal Control, covering its foundations, strengths, and limitations. We will survey the current theoretical understanding of SBOC, including recent results on convergence analysis and performance guarantees. We will also demonstrate the practical effectiveness of SBOC through compelling applications, particularly in challenging robotic control problems.

Autonomous Multi-Agent Systems in Transportation: Control, Learning, and Optimization Methods

Abstract: Emerging mobility systems are an example of Cyber-Physical Systems (CPSs) in which multiple autonomous agents (vehicles) interact with each other as well as with the infrastructure resources (road side units, traffic lights, etc). Control-theoretic and optimization methods provide a rich framework for managing these complex mixed-traffic socio-economic multi-agent systems. Given the complexity involved and the abundance of data now available, it is essential to integrate learning-based methods not only to design optimal controllers with safety guarantees, but to also gain an understanding of human driving behavior, as well as user preferences for the mobility options that intelligent transportation systems provide. The three objectives of the proposed tutorial session are: (1) Set the stage for emerging mobility systems consisting of both autonomous and human-driven vehicles in a mixed traffic environment by formulating basic optimal control problems for autonomous vehicles that seek to jointly optimize travel time, energy, and comfort while ensuring that safety constraints are always satisfied. (2) Present methods for solving the formulated problems using a combination of optimization techniques and Control Barrier Functions (CBFs) that provide safety guarantees, as well as state of the art learning-based methods to design effective controllers for mixed traffic transportation systems. (3) Address the societal issues accompanying emerging mobility systems, including new metrics that incorporate accessibility and fairness in a transportation network consisting of both autonomous and human-driven vehicles. The session will consist of two parts: three tutorial presentations followed by an open discussion with questions from the audience addressed to the three speakers.

Contraction Theory in Control, Optimization, and Learning

Abstract: Contraction theory is a powerful mathematical framework for analyzing convergence, robustness, and modularity of dynamical systems, optimization algorithms, and learning methods. Originating from the seminal works of Banach, Demidovich, Krasovski, Desoer, and Slotine, contraction theory provides a unifying set of concepts and tools to systematically study dynamical systems exhibiting exponential stability, incremental stability, and robustness to perturbations and uncertainties. This tutorial will introduce and survey the state-of-the-art in contraction theory, including theoretical foundations, computational methods, robustness properties, and applications to control, optimization, machine learning, and beyond.

Distributed Multi-Armed Bandits

Abstract: The goal of this tutorial session is to provide a clear and accessible introduction to multi-armed bandit problems and their recent extensions to distributed multi-agent settings over networks. We aim to explain the core concepts and challenges in networked bandit models, highlight their practical importance in areas such as defense, industry, and finance, and present recent advances in algorithms and theory. The session is especially designed for graduate students and researchers who are new to this topic. It will help them understand how distributed agents can make decisions under uncertainty while interacting with one another through limited communication. By the end of the tutorial, participants will gain both a solid understanding of the basic models and a sense of current research directions and open questions in this growing area.

Graphons in Systems and Control

Abstract: In this tutorial, we will introduce notions of structural systems theory, namely structural controllability (and its dual, structural observability) and structural stability, and show how to study them over random structures sampled from graphons.

IEEE CSS TC on Smart Cities Tutorial Session: Challenges and Opportunities for Control in Smart Cities

Abstract: Several urban areas are currently struggling with challenges such as grid congestion, (sustainable) energy and water procurement and distribution, environmental pollution, road traffic, cybersecurity, lack of accessibility and multi-dimensional social injustices. At the same time, we are witnessing significant advancements in new technologies such as the Internet of Things, AI and Big Data, autonomous driving, smart grid, smart water networks, and renewable energy technologies. These new technological trends continuously contribute to creating narratives around the concept of Smart Cities, where technologies are deployed to improve the wellbeing of citizens and the environment in a fair and sustainable manner. In line with Langdon Winner’s argument that (technological) “artifacts have politics”, now more than ever (control) engineers have the double responsibility of not only devising effective technological solutions but also ensuring that they are catering to the needs of society and the planet in terms of wellbeing, justice and sustainability in the first place. Against this background, the IEEE CSS Technical Committee on Smart Cities organizes a tutorial session to reason upon challenges and opportunities in a field that is growing in terms of technical results, interdisciplinarity and fascinating research areas. The session will feature four invited presentations on

• Smart mobility systems
• Smart power systems
• Smart water networks
• Transdisciplinarity

Objectives and Expected Outcome: Control systems can play an important role in shaping the future of Smart Cities. The expertise and innovative solutions of the community are crucial in addressing the complex challenges urban areas face today. This tutorial session provides an opportunity to learn about the latest advancements within exciting interdisciplinary areas revolving around Smart Cities, as well as to discuss practical implementations and explore new perspectives. By participating in this session, researchers in control will be better equipped to contribute to the development of Smart Cities that are not only technologically advanced, but also equitable, sustainable, and focused on the wellbeing of its citizens and the environment.