Uploaded image for project: 'GMAT'
  1. GMAT
  2. GMT-6227

Schur factorization method has issues converging with some matrices

    Details

    • Type: Bug
    • Status: Open (View Workflow)
    • Priority: P3
    • Resolution: Unresolved
    • Affects Version/s: R2018A
    • Fix Version/s: Someday
    • Component/s: Miscellaneous
    • Labels:
      None

      Description

      The Schur factorization method uses QR factorization to compute [Q, R] = qr(A) and
      Anew = Q*R. This cycle is repeated until the change in matrix A reaches a tolerance. At this point, A will represent the upper triangular matrix from Schur while the unitary matrix from Schur is calculated by multiplying each Q matrix used together.

      This should theoretically work for most matrices but some have issues converging for an unknown reason. Below is an example matrix that does not converge.
      A = [21 3 5 1 9 3 4 2
      4 18 5 9 2 3 4 1
      6 9 20 4 7 11 3 4
      3 8 2 32 9 7 7 2
      3 8 2 7 16 2 1 1
      8 10 2 3 2 30 7 6
      1 2 1 4 7 1 44 1
      9 10 2 3 6 1 1 25]
      It would be useful to find what is causing this bad convergance in order to see if it is an issue that can be fixed or should be trapped so the user knows ahead of time when a different matrix should be used.

        Gliffy Diagrams

          Attachments

            Activity

            There are no comments yet on this issue.

              People

              • Assignee:
                Unassigned
                Reporter:
                jraymond Joshua Raymond
                Participants:
              • Votes:
                0 Vote for this issue
                Watchers:
                1 Start watching this issue

                Dates

                • Created:
                  Updated: