Foundations of Reinforcement Learning and Control:Connections and New Perspectives
2026-08-03 • Machine Learning
Machine Learning
AI summaryⓘ
The authors explain that reinforcement learning and control theory are two related but different ways people try to teach machines to control systems they don't fully know about. They highlight how each field has developed unique methods and goals, which has created a gap between them. The authors introduce concepts like adaptive control and actor-critic algorithms and show a new way to combine these ideas using a movement control example. Their goal is to help experts from both fields understand each other better and work together more effectively.
reinforcement learningcontrol theoryadaptive controlactor-critic algorithmsdynamic programmingdynamical systemsfeedback controldata-driven decision makinglocomotion control
Authors
Claire Vernade, Onno Eberhard, Martha White, Florian Dörfler, Csaba Szepesvári, Miroslav Krstic, Michael Muehlebach
Abstract
Reinforcement learning and control theory are two adjacent scientific fields that focus on optimizing the controller of unknown dynamical systems using feedback. While both fields have common roots in dynamic programming, they have evolved with distinct methodologies, goals, and cultures. Despite decades of mutual influence, a significant gap persists between the two communities. This tutorial introduces adaptive control, actor-critic reinforcement algorithms, and a new way to combine these two paradigms for data-driven decision making on a classical locomotion control problem. Our aim is to provide a foundation for understanding the core differences between the two approaches and insights to help experts in each field better understand and engage with the tools and approaches of the other.