Show Notes

Decision-making is an essential part of everyday life and one of the main applications of data science is making the decision-making process easier.

However, mostly when data scientists build models, it’s to make a single decision. But in real life, decision-making is rarely that simple.

In this episode, Prof Warren Powell joins Dr Genevieve Hayes to discuss one way in which the decision-making process can become more complicated, in the form of sequential decision problems.

Guest Bio

Warren Powell is the co-founder and Chief Innovation Officer of Optimal Dynamics and a Professor Emeritus after retiring from Princeton, where he was a faculty member in the Department of Operations Research and Financial Engineering. He is also the author of Sequential Decision Analytics and Modelling and Reinforcement Learning and Stochastic Optimization.

Talking Points

  • What is a sequential decision problem?
  • Real-life examples of sequential decision problems and the disciplines in which they occur.
  • The four main classes of techniques for solving sequential decision problems.
  • How Warren’s approach to addressing sequential decision problems differs from the standard approach in this space.
  • The challenges of implementing sequential decision analysis techniques in practice.


podcast cover art
Value Driven Data Science
Episode 36: Sequential Decision Problems