Dr. Shalabh Bhatnagar

ACCS – CDAC Foundation Lecture 2017

Title: Efficient Algorithms for Optimization under Uncertainty

Abstract: In this talk, I will first present a general framework for optimization under uncertainty, also popularly called stochastic optimization. I will then discuss some of our work towards the design of stochastic approximation algorithms for solving such problems. It turns out that in many such algorithms the gradient/Hessian  estimators have biases that arise naturally. So far there had not been a satisfactory analysis of such algorithms with biases because all previous analyses showed convergence under the requirement that the bias vanishes asymptotically. It turns out that in the presence of bias, the dynamics of the underlying system becomes set-valued. I will present some of our work that shows the stability and convergence of these algorithms under set-valued dynamics. Our recent results on stability of set-valued stochastic approximations are the first and only known results in this direction. Finally, I will give an application of some of these algorithms with biases for the problem of traffic signal control in vehicular traffic networks.

Bio:

Prof. Shalabh Bhatnagar received Bachelors in Physics (Hons) from the University of Delhi in 1988. He received his Masters and Ph.D degrees in Electrical Engineering from the Indian Institute of Science, Bangalore in 1992 and 1997, respectively.

He was a Research Associate at the Institute for Systems Research, University of Maryland, College Park, during 1997 to 2000 and a Divisional Postdoctoral Fellow at the Free University, Amsterdam, during 2000 to 2001. He joined the Department of Computer Science and Automation at the Indian Institute of Science, Bangalore in December 2001, where he is now a Professor. He has also held visiting faculty positions at the Indian Institute of Technology, Delhi and the University of Alberta, Canada.

Prof. Bhatnagar’s interests are in simulation based stochastic optimization, stochastic control and reinforcement learning. He has authored or co-authored more than 120 research articles in various journals and conferences. He is also the coauthor of a book with title `Stochastic Recursive Algorithms for Optimization: Simultaneous Perturbation Methods’, published by Springer in 2013. He is a Senior Associate of the International Center for Theoretical Physics (ICTP), Italy, a Fellow of the Indian National Academy of Engineering and a Fellow of the Institution of Electronics and Telecommunication Engineers. He is the winner of ACCS-CDAC Foundation Award 2017.