Skip to content
University of Maryland LibrariesDigital Repository at the University of Maryland
    • Login
    View Item 
    •   DRUM
    • A. James Clark School of Engineering
    • Institute for Systems Research Technical Reports
    • View Item
    •   DRUM
    • A. James Clark School of Engineering
    • Institute for Systems Research Technical Reports
    • View Item
    JavaScript is disabled for your browser. Some features of this site may not work without it.

    Distributed subgradient method under random communication topology - the e

    Thumbnail
    View/Open
    IMatei_JBaras_MASM_edited.pdf (606.7Kb)
    No. of downloads: 626

    Date
    2009-09
    Author
    Matei, Ion
    Baras, John
    Advisor
    Baras, John
    Metadata
    Show full item record
    Abstract
    In this note we study the performance metrics (rate of convergence and guaranteed region of convergence) of a multi-agent subgradient method for optimizing a sum of convex functions. We assume that the agents exchange information according to a communication topology modeled as a random graph, independent of other time instances. Under a strong convexity type of assumption, we express the performance metrics directly as functions of the estimates of the optimal decision vector. We emphasize how the probability distribution of the random graph affects the upper bounds on the performance metrics. This provide a guide for tuning the parameters of the communication protocol such that good performance of the multi-agent subgradient method is ensured. We perform the tuning of the protocol parameters for two communication scenarios. In the first scenario, we assume a randomized scheme for link activation with no-error transmissions while in the second scenario we use a pre-established order of transmissions but we consider the interference effect. Both these scenarios are applied on a small world type of topology.
    URI
    http://hdl.handle.net/1903/9441
    Collections
    • Institute for Systems Research Technical Reports

    DRUM is brought to you by the University of Maryland Libraries
    University of Maryland, College Park, MD 20742-7011 (301)314-1328.
    Please send us your comments.
    Web Accessibility
     

     

    Browse

    All of DRUMCommunities & CollectionsBy Issue DateAuthorsTitlesSubjectsThis CollectionBy Issue DateAuthorsTitlesSubjects

    My Account

    LoginRegister
    Pages
    About DRUMAbout Download Statistics

    DRUM is brought to you by the University of Maryland Libraries
    University of Maryland, College Park, MD 20742-7011 (301)314-1328.
    Please send us your comments.
    Web Accessibility