eduzhai > Applied Sciences > Engineering >

Distributed Personalized Gradient Tracking with Convex Parametric Models

  • king
  • (0) Download
  • 20210506
  • Save

... pages left unread,continue reading

Document pages: 27 pages

Abstract: We present a distributed optimization algorithm for solving onlinepersonalized optimization problems over a network of computing andcommunicating nodes, each of which linked to a specific user. The localobjective functions are assumed to have a composite structure and to consist ofa known time-varying (engineering) part and an unknown (user-specific) part.Regarding the unknown part, it is assumed to have a known parametric (e.g.,quadratic) structure a priori, whose parameters are to be learned along withthe evolution of the algorithm. The algorithm is composed of two intertwinedcomponents: (i) a dynamic gradient tracking scheme for finding local solutionestimates and (ii) a recursive least squares scheme for estimating the unknownparameters via user s noisy feedback on the local solution estimates. Thealgorithm is shown to exhibit a bounded regret under suitable assumptions.Finally, a numerical example corroborates the theoretical analysis.

Please select stars to rate!

         

0 comments Sign in to leave a comment.

    Data loading, please wait...
×