Speaker: Jakob Kotas, Applied Mathematics at the University of Washington–Seattle

Title: Response-guided dosing

Abstract: Within the broad field of personalized medicine, there has been a recent surge of clinical interest in the idea of response-guided dosing. Roughly speaking, the goal is to devise strategies that administer the right dose to the right patient at the right time. We will present stochastic models that attempt to formalize such optimal dosing problems. Theoretical results about the structure of optimal dosing strategies and associated solution methods rooted in convex optimization, stochastic dynamic programming, robust optimization, and Bayesian learning will be described. Computational results on rheumatoid arthritis will be discussed.