Control of large-scale complex networks has received a lot of attention in recent years due to their ubiquitous applications in power grids, robotics, social networks, and transportation. In this talk, I shall discuss two modes of network control for fixed communication topology, namely, control using edges and control using nodes. Network control via manipulation of edge weights (interaction rates) yields a bilinear model where the current state of the system gets multiplied with the control input to evolve to the next state. I will describe exact algebraic and graph-theoretic conditions for controllability of such bilinear networks which solely uses information about their communication structure. These "structural" properties hold for almost all choices of system parameters. In the second half, I will discuss the alternate form of network control via injection of external control decisions at nodes. This will be presented in the context of multi-agent consensus problems for two scenarios: under known topologies (teams) and unknown topologies (games). Decentralized linear dynamic feedback control schemes are developed that guarantee optimal consensus.
Supratim Ghosh is currently working as a Senior Research Fellow at the University of Michigan, Ann Arbor on a NASA-funded project related to fault diagnosis in complex systems. Prior to this, he spent four years in the Engineering Systems and Design pillar at the Singapore University of Technology and Design: first three years as a Postdoctoral Research Fellow and then a year as a Lecturer (Teaching-track faculty). He received his undergraduate degree in Instrumentation and Control Engineering from National Institute of Technology, Jalandhar, India. He then went on to receive his M.S. in Electrical Engineering, M.A. in Mathematics, and Ph.D. in Electrical Engineering all from the Pennsylvania State University, USA. His current research interests include decentralized and optimal control, distributed optimization, game theory, and their applications to problems in networks, power systems, and robotics.