Weekly seminar in sequential decision problems for graduate students and analytics professionals…
Sequential decision problems are universal - they arise in every human process. However, the academic community teaches this field in a hopelessly fragmented way: dynamic programming, stochastic programming, stochastic search, optimal control, simulation-optimization, multiarmed bandit problems and (especially) "reinforcement learning."
All of these fields work on "sequential decision problems."
My new book is the first to pull all these fields together into a single, unified framework. I organize every method used for making decisions into four classes of policies.
This material is not being taught in universities today. I have designed a weekly seminar where graduate students and analytics professionals can teach the material to themselves. Just go to
https://tinyurl.com/RLSOcourses/
and scroll down to "Weekly seminar series" where I provide an outline of material that a group of students or professionals can follow to teach the material to themselves from my recent book https://tinyurl.com/RLandSO/. Please also see the companion book at https://tinyurl.com/sdamodeling/ which uses a teach-by-example style.
Additional material is available at https://tinyurl.com/sdalinks. ------------------------------
Warrren Powell
Chief analytics officer, Optimal Dynamics
Professor Emeritus, Princeton University
http://www.castlelab.princeton.edu/------------------------------