When:
Tuesday, June 26, 2018
11:00 AM - 12:00 PM CT
Where: Technological Institute, F160, 2145 Sheridan Road, Evanston, IL 60208 map it
Audience: Faculty/Staff - Student - Public - Post Docs/Docs - Graduate Students
Contact:
Cristian Pennington
(847) 491-3645
Group: Physics and Astronomy Complex Systems Seminars
Category: Academic
Dr. Ahmad F. Taha, Univ. of Texas at San Antonio
Title: A Generalized Approach to the Sensor & Control Nodes Selection/Placement Problems in Dynamic Networks
Abstract:
Transportation networks, power grids, and many cyber-physical systems are naturally or artificially assembled into networks—intertwined, dynamic networks that evolve nonlinearly in time and space. A defining feature of these networks is the prevalence and ubiquity of sensors and control nodes (S&C): sensors collecting data while ensuring network observability, and control nodes driving networks to stability. The selection/placement of S&C is one of the main steps in the network design and control. Broadly speaking, the S&C selection/placement problems can be described as the joint routines of: (a) Finding S&C selection that satisfy system constraints such as closed-loop stability and bounded estimation error, (b) Optimizing system-theoretic metrics such as linear quadratic regulation, Kalman filtering, robust control/estimation, while (c) Designing localized control/estimation routines for each node.
This talk presents a generalized approach to solve the S&C selection problem considering a wide range of network dynamics and control-theoretic metrics. In particular, the approach considers linear and nonlinear dynamic networks, under the presence or absence of disturbances, as well as robust control/estimation metrics. The approach is based on classifying network nonlinearities, and using mixed-integer nonconvex programming formulations. The generalized formulation can then be solved via a plethora of methods ranging from convex relaxations/approximations, binary search algorithms, heuristics, cutting plane methods, and sub-optimal branch and bound routines. The scalability of the approach is discussed, and applications to transportation networks and power grids are showcased.
Host: Adilson Motter