BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Feedback-based online algorithms for time-varying optimization: th
 eory and applications in power systems - Emiliano Dall'anese (University o
 f Colorado)
DTSTART:20190110T113000Z
DTEND:20190110T123000Z
UID:TALK116797@talks.cam.ac.uk
CONTACT:INI IT
DESCRIPTION:The talk focuses on the synthesis and analysis of online algor
 ithmic solutions to control systems or networked systems based on performa
 nce objectives and engineering constraints that may evolve over time. Part
 icular emphasis is given to applications in power systems operations and c
 ontrol. The time-varying optimization formalism is leveraged to model opti
 mal operational trajectories of the systems\, as well as explicit local an
 d network-level constraints. The design of the algorithms then capitalizes
  on an online implementation of primal-dual projected-gradient methods\; t
 he gradient steps are\, however\, suitably modified to accommodate actiona
 ble feedback in the form of measurements from the network -- hence\, the t
 erm feedback-based online optimization. By virtue of this approach\, the r
 esultant running algorithms can cope with model mismatches in the algebrai
 c representation of the system states and outputs\, they avoid pervasive m
 easurements of exogenous inputs\, and they naturally lend themselves to a 
 distributed implementation. Under suitable assumptions\, Q-linear converge
 nce to optimal solutions of a time-varying convex problem is shown. On the
  other hand\, under a generalization of the Mangasarian-Fromovitz constrai
 nt qualification\, sufficient conditions are derived for the running algor
 ithm to track a Karush-Kuhn-Tucker point of a time-varying nonconvex probl
 em. Examples of applications in power systems will be provided. <br> <br> 
 Joint work with: A. Simonetto\, Y. Tang\, A. Bernstein\, and S. Low.
LOCATION:Seminar Room 1\, Newton Institute
END:VEVENT
END:VCALENDAR