Order-reducing conjugate gradients versus block AOR for constrained least-squares problems (Q808173)

From MaRDI portal
scientific article
Language Label Description Also known as
English
Order-reducing conjugate gradients versus block AOR for constrained least-squares problems
scientific article

    Statements

    Order-reducing conjugate gradients versus block AOR for constrained least-squares problems (English)
    0 references
    0 references
    1991
    0 references
    The convergence properties of two iterative methods for solving equality- constrained least-squares problems are compared. The first one is a modified conjugate gradient algorithm. The second one is the block accelerated overrelaxation, a two-parameter generalization of block successive overrelaxation. It is proved that the first method converges faster than the second one. Numerical experiments support the analysis.
    0 references
    0 references
    convergence
    0 references
    iterative methods
    0 references
    least-squares problems
    0 references
    conjugate gradient algorithm
    0 references
    block accelerated overrelaxation
    0 references
    successive overrelaxation
    0 references
    Numerical experiments
    0 references
    0 references
    0 references