AE Presents: "Towards a Scalable Framework for Analysis and Control of Networked Systems"

Mon Feb 05 2018 11:00 AM to 12:00 PM
Montgomery Knight 317
A talk by Cal Tech's Dr. James Anderson

You are invited to hear

"Towards a Scalable Framework for Analysis
and Control of Networked Systems"

a talk by

Dr. James Anderson

Senior Postdoctoral Scholar
Department of Computing and Mathematical Sciences
California Institute of Technology

About the Talk

We live in the network era. From emerging technologies such as smart energy grids and ride sharing platforms, to well-established systems such as the Internet and transportation networks, our day-to-day lives depend on engineers understanding how these networks function – and how they fail. In the classical setting, designing controllers that mitigate the effects of disturbances and uncertainty requires a mathematical model of the system to be controlled. However, in the network setting this is problematic as the model complexity quickly renders the control design task intractable, and moreover, even simulation and analysis can be problematic. In this talk I describe two approaches to attacking this problem. The first is based on the idea of constructing a reduced-order (less complex) model of the network that provides a “good” approximation of the original network but is of a size that is not computationally prohibitive. I outline recent work on how this can be achieved in a way that has favorable computation constraints, and which, unlike traditional approaches, maintains the structure of the underlying network. In the second part of the talk I introduce the recently formulated System Level Synthesis (SLS) paradigm for distributed control. The SLS approach uses the idea of locality (or virtual locality) in space and time in order to decouple the control design problem; in short, it allows for O(1) synthesis and implementation complexity relative to the dimension of the networked system. I illustrate these ideas with simple engineering examples taken from electrical and aerospace engineering and conclude by sketching out some future challenges for this exciting field.

About the Speaker

James Anderson is a senior postdoctoral scholar in the Department of Computing and Mathematical Sciences at the California Institute of Technology. Before his position at Caltech he held a Junior Research Fellowship at St John’s College, Oxford and was a member of the Control Group based in the Department of Engineering Science at the University of Oxford. He obtained the BSc and MSc degrees from the School of Systems Engineering at the University of Reading, UK, and a DPhil (PhD) in Engineering Science from the University of Oxford in 2012. His research interests broadly span design, analysis, optimization, and validation of networked systems ranging from cyber-physical to biological. He is also a lead author of the sum-of-squares programming toolbox SOSTOOLS.

Sponsored by Prof. Panagiotis Tsiotras


Montgomery Knight 317