Choices Are Not Independent: Stackelberg Security Games with Nested Quantal Response Models

Oct 5, 2021 04:00 PM Singapore (Registration will open at 03:50 PM.)

Join Zoom Meeting:
https://sutd-edu.zoom.us/j/99404847706?pwd=QUM1bzJrS2k5ektIK3E5L1R5VVl1UT09

Meeting ID: 994 0484 7706
Passcode: S?30JdR0

Abstract

We introduce the use of the nested logit model in Stackelberg security games which addresses a shortcoming of the classical multinomial logit model used in prior work. We present tractable approximation of the resulting equilibrium problem with nested quantal response (QR) adversary. We do so by deriving an interesting property of the equilibrium problem, namely a loosely coupled split into nested problems that mirrors the nested decision making by the adversary in the nested QR model. We show that each separate nested problem can be approximated efficiently and that the loosely coupled overall problem can be solved approximately by formulating it as a discretized version of a continuous dynamic program. Finally, we conduct experiments that show the scalability and parallelizability of our approach, as well as advantages of the nested QR model. Our work can also tackle previously unaddressed constrained pricing problems in the product pricing literature.

About the Speaker

Tien Mai is an assistant professor at School of Computing and Information Systems, Singapore Management University. He holds a Ph.D. in Operations Research from University of Montreal. Prior to SMU, he was a postdoc scholar at Singapore-MIT Alliance for Research and Technologies and Massachusetts Institute of Technologies. His work has been recognized by the INFORM-TSL Dissertation Prize for best doctoral dissertation in the area of transportation science and logistics, the Eric Pas Dissertation Prize for best doctoral dissertation in travel behaviour research and a Best Paper Award from the European Association for Research in Transportation. His research interests include data-driven optimization, discrete choice modeling and imitation learning, with applications in transportation modeling, revenue/workforce management and security game.

For more information about the ESD Seminar, please email esd_invite@sutd.edu.sg