Bi-Level Dynamic Optimization with Feedback

Andrey Bernstein, Emiliano Dall-Anese

Research output: Contribution to conferencePaper


This paper considers a bi-level real-time algorithmic framework for networked systems, consisting of several local controllers and a central controller. The central controller issues setpoints to the local controllers to optimize their operational objectives while satisfying system-wide constraints. In this context, the paper develops an online algorithm for tracking the optimal solution of the underlying dynamic optimization problem. The design of the algorithm is based on a projected-gradient method, suitably modified to accommodate appropriate measurements (i.e., feedback). Optimality claims are established in terms of the dynamic regret of the algorithm; the latter is a natural performance criterion in nonstationary environments associated with real-time control problems. Finally, the application of the algorithm to real-time control of power setpoints in an electrical grid is illustrated.
Original languageAmerican English
Number of pages5
StatePublished - 2018
Event2017 IEEE Global Conference on Signal and Information Processing (GlobalSIP) - Montreal, Canada
Duration: 14 Nov 201716 Nov 2017


Conference2017 IEEE Global Conference on Signal and Information Processing (GlobalSIP)
CityMontreal, Canada

NREL Publication Number

  • NREL/CP-5D00-71825


  • government
  • heuristic algorithms
  • linear programming
  • optimization
  • power system dynamics
  • real-time systems
  • signal processing algorithms


Dive into the research topics of 'Bi-Level Dynamic Optimization with Feedback'. Together they form a unique fingerprint.

Cite this