Sequential search is a central problem in decision making under uncertainty. In the classical search problem, a decision maker (DM) receives a sequence of independent and identically distributed observations, each which comes at a cost, from a possibly unknown distribution. After each observation, the DM has the option to stop the search, receiving the highest observation so far as the payout from this process. The goal is to identify a policy which will maximize the DM’s expected payout from the process.
Variations on this model are pervasive in economics, decision analysis, operations research, computer science, statistics, and marketing. Search models have been used to describe a variety of situations ranging from the job market, real estate market, pharmaceutical R&D, capital investments, to technology adoption. In particular, the applications of this theory in economics culminated in a Nobel prize being awarded for the study of “markets with search frictions” in 2010. In this talk, I will provide a brief overview on the basics of this theory as well as some of the recent advances coming from operations research, with an emphasis on Zorc and Tsetlin (2020), Zorc et al. (2023), and Baucells and Zorc (2023). Particularly, the focus will be on situations where the underlying distribution in not known but has to be inferred from observations while searching.
References:
Zorc, S., & Tsetlin, I. (2020). Deadlines, offer timing, and the search for alternatives. Operations Research, 68(3), 927-948.
Zorc, S., Tsetlin, I., Hasija, S., & Chick, S. E. (2023). The when and how of delegated search. Operations Research.
Baucells, M., & Zorc, S. (2023). Search in the dark: The normal case (working paper).
