Planning ahead: using Markov decision processes to optimize your life

Abstract:

Have you ever found yourself in a situation where there were multiple courses of action with different potential outcomes? How do you decide what to do? In this talk, Eric Cotner will discuss Markov decision processes (MDP's); a quantitative framework for determining an optimal sequence of actions. We'll learn how to set up an MDP, what kinds of situations they can be applied to, some of the math behind it, and finally, an easy way to solve simple MDP's using the python package `pymdptoolbox`. If you're interested in applying optimal decision making to your life or work, you won't want to miss this!

Speaker Bio:

Eric Cotner is a former theoretical physicist who studied exotic dark matter candidates such as Q-balls and black holes during his graduate school years. More recently, he has turned his attention to data science; specifically prescriptive analytics. He works as a data scientist at American Tire Distributors in Huntersville, where he focuses on supply chain and logistics problems such as optimizing delivery routes and modeling inter-warehouse product flow. In his free time, he enjoys hiking, astronomy, and scuba diving. He holds a PhD/Masters/Bachelors in Physics from UCLA/UCLA/UCSD.
Venue Sponsor:

A huge thanks to UNC Charlotte - SCHOOL OF DATA SCIENCE https://datascience.uncc.edu/ for hosting us!

Contact:

For questions about this meetup, please message through the Meetup site or email contact@pydatacharlotte.com