Knowing When to Stop
Speaker: Theodre Hill, Georgia Institute of Technology
Abstract: In many basic processes in science (and the rest of life) there is an element of chance involved, and a crucial problem is deciding when to stop. The process could be waiting to buy or sell Google stocks, proofreading a paper or debugging a large software program, deciding when to switch to a new medication, updating eBay auction bids, or interviewing for a new secretary (or spouse). At some point you need to stop, and your objective is to do it in a way that optimizes your reward (e.g. maximum profit or satisfaction, minimum cost or errors). The mathematical theory of optimal stopping, including Secretary Problems (also known in the literature as Marriage, Dowry, or Best-Choice Problems), has a long and colorful history, complete with excellent rules of thumb, counterintuitive surprises, colorful paradoxes, and famous unsolved problems. The elegant and unexpected solution to the classical “no-information” Secretary Problem will be reviewed, along with several game-theoretic extensions, analogs for “full-information” and “partial-information” stopping, and several basic open problems.
Room Reservation Information
Room Number: 113 McAllister
Time: 1:25pm - 2:25pm