Risk Aware Online Optimization
Speaker (s):
William Haskell
Assistant Professor
Department of Industrial & Systems Engineering
National University of Singapore
|
|
Date:
Time:
Venue:
|
|
April 15, 2016, Friday
2:00pm - 3:30pm
Meeting Room 4.4, Level 4
School of Information Systems
Singapore Management University
80 Stamford Road
Singapore 178902
We look forward to seeing you at this research seminar.
|
|
About the Talk
Classical online optimization and machine learning methods focus on minimizing expected cost which is a "risk-neutral" performance objective. Given the significant interest in risk mitigation in operations research, we seek to develop online optimization and machine learning tools for "risk-aware" performance objectives. In this talk, we develop a class of data-driven primal-dual algorithms that can solve two classes of risk-aware optimization problems: stochastic dominance constrained optimization and coherent risk minimization.
In both cases, we derive bounds on the optimality gap as a function of the number of observed data points. Preliminary numerical experiments show that these algorithms outperform their theoretical error guarantees.
About the Speaker
William Haskell completed his B.S. Mathematics degree at the University of Massachusetts Amherst in Spring 2006, then went on to complete his M.S. Mathematics and Ph.D. Operations Research degrees at the University of California Berkeley in 2011. He worked as a visiting assistant professor in the Industrial and Systems Engineering department at the University of Southern California starting in 2011, until he joined the ISE department at the National University of Singapore in Fall 2014. Will's current work emphasizes risk-aware decision-making, large-scale optimization, and dynamic programming.
