Rafael L. Beirigo

Laboratório Nacional de Computação Científica

Papers

Total Citations

H-Index

About

Rafael L. Beirigo is a researcher specializing in stochastic control theory and reinforcement learning, with a particular focus on Markov jump linear systems (MJLS). His work bridges the gap between classical control theory and modern machine learning approaches, addressing fundamental challenges in systems with random parameter variations. Beirigo's most significant contribution is the development of an online temporal differences (TD) algorithm for discrete-time MJLS, inspired by reinforcement learning concepts. This pioneering work, published in 2018 and garnering 8 citations, proposes a method that can simultaneously apply and refine control policies in real-time, representing a novel fusion of adaptive control and learning-based techniques. His subsequent research in 2017 addressed the practical challenge of controlling MJLS when transition probabilities are unknown, proposing a count-based quadratic control approach that achieves optimal performance despite parametric uncertainty. While his citation counts reflect the emerging nature of this specialized field, Beirigo's work has established important theoretical foundations for integrating reinforcement learning principles with traditional control of stochastic systems, offering promising pathways for applications in robotics, autonomous systems, and adaptive control where system dynamics are subject to random jumps.