Computational Stochastic Optimization and Learning

As of Sept 1, 2020, I have retired from Princeton University.  I am now the Chief Analytics Officer of Optimal Dynamics which licensed my library for truckload trucking and dynamic resource allocation.

I am pleased to announce that I recently won the Robert Herman Lifetime Achievement Award from the Society for Transportation Science and Logistics within Informs. I prepared for the award that summarized my life path that led to my universal framework for sequential decision problems.

Warren Powell
Professor Emeritus, Princeton University


CASTLE works to advance the development of modern analytics for solving a wide range of applications that involve decisions under uncertainty.

New book! Reinforcement Learning and Stochastic Optimization: A unified framework for sequential decisions.

New resource page for sequential decision analytics.

  • We have initiated a new field we are calling Sequential Decision Analytics which offers a universal framework for sequential decision problems under uncertainty, which unifies 15 different communities that all deal with decisions under uncertainty.
  • Teaching – We now teach this framework to both undergraduates and graduate students across a range of departments.
  • We have created a new class of policy based on the idea of parameterizing a deterministic optimization model for solving complex stochastic problems.  We call this “parametric cost function approximations.”
  • What is AI? – My attempt at describing “artificial intelligence.”
  • What is reinforcement learning? – My attempt at describing “reinforcement learning.”

Laboratories and Areas of Focus

Surrounding the core activities in methodology are laboratories focusing on major areas of application.

  • PENSA – The Princeton Laboratory for Energy Systems Analysis – PENSA addresses a variety of stochastic optimization problems in energy systems, including energy storage, stochastic unit commitment, bidding, and pricing.  We document our work on energy storage.
  • Transportation and logistics Laboratory – Our work in transportation and logistics dates to 1981, and spans stochastic fleet management in trucking, rail and air, real-time dispatching, routing and scheduling, and spare parts management, to name a few. Our work has been adopted by a broad cross section of the industry.
  • Optimal learning – This research addresses the challenges of collecting information, when information (observations, simulations, laboratory and field experiments) are expensive.
  • Health sciences – Projects in health have included drug discovery, drug delivery, blood management, dosage decisions, personal health, and health policy.

I hope you find the material interesting, and perhaps useful. If you have any questions, please contact me.

Warren Powell