Learning and Optimal Control in Dynamical Systems

Program Description

Learning to act optimally in dynamical environments, where decisions shape future observations and uncertainty, is central to safe, data-efficient autonomous systems. From stabilizing and decarbonizing energy grids, to adaptive treatment in chronic care, to robotics and space exploration, society increasingly relies on dynamic decision-making systems that must trade off information acquisition and performance under safety, resource, and data constraints. 

Concurrently, core ideas and tools for these challenges are advancing across several vibrant communities, including control and systems theory, statistical learning, online optimization and reinforcement learning, and formal models of learning and action. These approaches often adopt different assumptions, objectives, and techniques.

The P1 program, Learning and Optimal Control in Dynamical Systems, aims to build a sustained community to create a common language, surface shared problem structure, and enable systematic exchange of methods, open problems, and application-driven challenges. Additionally, the program seeks to catalyze theoretical insights and algorithmic advances, accelerating progress toward principled and scalable approaches for learning and optimal control in dynamical systems. 

Beyond knowledge exchange, the ambition is to foster enduring collaborations across Denmark and the Nordics, seed joint research agendas, support cross-institutional teams, and enable coordinated efforts, including collaborative grant proposals and invited/special sessions at major conferences.

As part of the program activities, two workshops will be held focusing on (1) establishing a coherent community around learning and optimal control in dynamical systems by bringing together researchers to exchange insights, identify shared technical bottlenecks and evaluation gaps, and stimulate the development of benchmarks and reusable methodologies grounded in application-driven challenges, and (2) building on the insights gained from workshop 1, shifting emphasis toward consolidation and sustainability, promoting new and persistent joint research directions with a forward-looking perspective, identifying the most pressing open problems.