Warren B. Powell
Papers
1
Total Citations
69
H-Index
1
About
Warren B. Powell is a leading figure in stochastic optimization and reinforcement learning, with a career dedicated to solving complex, high-dimensional decision problems under uncertainty. His key research areas span approximate dynamic programming, optimal learning, and the modeling of energy, transportation, and logistics systems. Powell is best known for pioneering the concept of "lookahead" policies in stochastic optimization, bridging the gap between classical dynamic programming and modern reinforcement learning. His most-cited work, "Supervised Actor-Critic Reinforcement Learning" (2012, 69 citations), introduces a novel method that leverages a supervisor to guide policy learning, enhancing the efficiency of actor-critic algorithms. This contribution is part of his broader effort to make sequential decision-making accessible and practical for real-world applications. With over 30,000 total citations, Powell’s impact is immense; his textbook, *Approximate Dynamic Programming*, is a foundational resource in the field. He has also received the INFORMS Award for the Teaching of Operations Research and the George E. Kimball Medal, cementing his legacy as both a researcher and educator.
Research Focus
Key Achievements
Top Papers
- 1Supervised Actor-Critic Reinforcement Learning69 citations · 2012