Title: Inner and outer-approximation algorithms for computing Nash equilibria in a family of mathematical-programming games
Speaker: Sriram Sankaranarayanan
Date, Day, Time: 13th April (Thursday), 9:30 am.
Venue: Seminar hall, IEOR Building, Second floor.
Abstract: This talk introduces a family of mathematical-programming games called reciprocally bilinear games (RBG). This includes integer programming games (IPG), Nash-Among-Stackelberg-Players (NASP) games among others. We show that in many cases, the problem of deciding if this game has a mixed-strategy Nash equilibrium is Sigma-2-p hard if at least one-player has an unbounded feasible set. We also show that the problem of deciding if this game has a pure-strategy Nash equilibrium is Sigma-2-p hard even if every player has a bounded feasible set. We then provide two finite algorithms, an inner approximation algorithm (valid for IPG and NASPs) and an outer approximation algorithm (valid for all RBGs) to compute a mixed-strategy Nash equilibrium or decide its non-existence.
Bio of the speaker: Prof. Sankaranarayanan is an Assistant Professor at IIM Ahmedabad in the area of Production and Quantitative Methods. His research interest lies in solving game-theoretic and optimisation problems that include integer variables. In particular, he has worked on mixed-integer linear programming, complementarity problems and mixed-integer bilevel programming. Apart from proving structural results and developing algorithms to solve these problems, he is also interested in using these methods for real-life problems which are of social interest. He has worked on using tools from optimisation to analyse energy-market policies, with a particular interest to combat climate change.
Host: Prof. Avinash Bhardwaj.