• Find People
  • Campus Map
  • PiratePort
  • A-Z
    • About
    • Submit
    • Browse
    • Login
    View Item 
    •   ScholarShip Home
    • Academic Affairs
    • Thomas Harriot College of Arts and Sciences
    • Mathematics
    • View Item
    •   ScholarShip Home
    • Academic Affairs
    • Thomas Harriot College of Arts and Sciences
    • Mathematics
    • View Item
    JavaScript is disabled for your browser. Some features of this site may not work without it.

    Browse

    All of The ScholarShipCommunities & CollectionsDateAuthorsTitlesSubjectsTypeDate SubmittedThis CollectionDateAuthorsTitlesSubjectsTypeDate Submitted

    My Account

    Login

    Statistics

    View Google Analytics Statistics

    Eigenvalues for Sums of Hermitian Matrices

    Thumbnail
    View/ Open
    Taylor_ecu_0600O_11437.pdf (440.2Kb)

    Show full item record
    Author
    Taylor, James M.
    Abstract
    In this thesis we explore how the eigenvalues of nxn Hermitian matrices A,B relate to the eigenvalues of their sum C=A+B. We mainly focus on inequalities bounding sums of r eigenvalues for C by sums of r eigenvalues for A with r eigenvalues for B, for some r less than n.     We begin by using linear algebra to establish some classical results, including a result by R.C. Thompson that allows us to reformulate the eigenvalue problem in terms of nonempty intersections in the Grassmannian manifold of r-planes in complex n-dimensional space. In particular, every nonempty triple intersection of Schubert varieties in a Grassmannian yields an eigenvalue inequality. Such nonempty intersections correspond to nonzero cup products in the cohomology ring of the Grassmannian, and consequently, to nonzero Littlewood-Richardson coefficients. The Littlewood-Richardson rules provide us with an efficient method of detecting when these coefficients are nonzero, and hence of finding eigenvalue inequalities which necessarily hold for all nxn Hermitian matrices A,B,C=A+B.    For the remainder of this thesis, we turn our attention to particular inequalities of the above form that Alfred Horn conjectured would completely determine the possible eigenvalues of A,B,C=A+B. Horn's conjecture, formulated in 1962, was resolved in the affirmative during the late 1990's in celebrated work of A. Knutson and T. Tao, building on results of A. Klyachko and others. We will develop the connection between Horn's inequalities and the earlier parts of this thesis. In particular, we will see that each Horn inequality corresponds to a nonzero cup product that lies in the top degree cohomology group of the Grassmannian.     An alternate formulation of Horn's Theorem shows that indices yield a Horn inequality if and only if certain associated partitions occur as the eigenvalues for some rxr Hermitian matrices A, B, C=A+B. We will prove that when r=n-2 there are necessarily diagonal rxr matrices satisfying this condition.  
    URI
    http://hdl.handle.net/10342/4954
    Subject
     Mathematics; Hermitian matrices; Horn inequalities 
    Date
    2015
    Citation:
    APA:
    Taylor, James M.. (January 2015). Eigenvalues for Sums of Hermitian Matrices (Master's Thesis, East Carolina University). Retrieved from the Scholarship. (http://hdl.handle.net/10342/4954.)

    Display/Hide MLA, Chicago and APA citation formats.

    MLA:
    Taylor, James M.. Eigenvalues for Sums of Hermitian Matrices. Master's Thesis. East Carolina University, January 2015. The Scholarship. http://hdl.handle.net/10342/4954. April 20, 2021.
    Chicago:
    Taylor, James M., “Eigenvalues for Sums of Hermitian Matrices” (Master's Thesis., East Carolina University, January 2015).
    AMA:
    Taylor, James M.. Eigenvalues for Sums of Hermitian Matrices [Master's Thesis]. Greenville, NC: East Carolina University; January 2015.
    Collections
    • Master's Theses
    • Mathematics
    Publisher
    East Carolina University

    xmlui.ArtifactBrowser.ItemViewer.elsevier_entitlement

    East Carolina University has created ScholarShip, a digital archive for the scholarly output of the ECU community.

    • About
    • Contact Us
    • Send Feedback