Orthogonal Diagonalization - Linear Algebra

  1. If P=12[[1,1],[1,−1]]P = \frac{1}{\sqrt{2}}[[1, 1], [1, -1]] and D=[[4,0],[0,0]]D = [[4, 0], [0, 0]] (rows), what is A=PDPTA = PDP^T?
    [[2,2],[2,2]][[2, 2], [2, 2]] (rows). It equals 4uuT4\mathbf{u}\mathbf{u}^T with u=12(1,1)\mathbf{u} = \frac{1}{\sqrt{2}}(1, 1).
  2. Why can P−1P^{-1} be replaced by PTP^T in an orthogonal diagonalization A=PDP−1A = PDP^{-1}?
    The columns of PP are orthonormal, so PTP=IP^TP = I.
  3. Which unit eigenvector with positive entries of [[5,2],[2,2]][[5, 2], [2, 2]] (rows) belongs to eigenvalue 66?
    (25,15)(\frac{2}{\sqrt{5}}, \frac{1}{\sqrt{5}}). A−6I=[[−1,2],[2,−4]]A - 6I = [[-1, 2], [2, -4]] has null space spanned by (2,1)(2, 1).
  4. A 3×33 \times 3 symmetric matrix has eigenvalues 2,2,52, 2, 5. How many orthonormal eigenvectors for λ=2\lambda = 2 go into PP?
    22. For a symmetric matrix the eigenspace dimension equals the multiplicity of the eigenvalue.
  5. Which unit eigenvector with positive entries of [[2,1],[1,2]][[2, 1], [1, 2]] (rows) belongs to eigenvalue 33?
    (12,12)(\frac{1}{\sqrt{2}}, \frac{1}{\sqrt{2}}). A(1,1)=(3,3)A(1, 1) = (3, 3), then normalize.
  6. How does orthogonal diagonalization differ from ordinary diagonalization A=PDP−1A = PDP^{-1}?
    The columns of PP must be orthonormal eigenvectors, so PP is orthogonal and P−1=PTP^{-1} = P^T.
  7. When orthogonally diagonalizing a symmetric matrix, what extra step is needed for an eigenspace of dimension 2 or more?
    Apply Gram-Schmidt within that eigenspace to get an orthonormal basis of it, then use those vectors as columns of PP.
  8. What are the eigenvalues placed in DD when orthogonally diagonalizing [[2,1],[1,2]][[2, 1], [1, 2]] (rows)?
    33 and 11. The trace is 44 and the determinant is 33.
  9. With eigenvalues in decreasing order, what is DD in an orthogonal diagonalization of [[0,2],[2,0]][[0, 2], [2, 0]] (rows)?
    [[2,0],[0,−2]][[2, 0], [0, -2]] (rows). The eigenvalues are ±2\pm 2, with eigenvectors along (1,1)(1, 1) and (1,−1)(1, -1).
  10. Why must the eigenvector columns of PP be unit vectors in A=PDPTA = PDP^T?
    With orthogonal but non-unit columns, PTPP^TP is diagonal but not II, so PTP^T would not be P−1P^{-1}.
  11. What are PP and DD in an orthogonal diagonalization of the diagonal matrix [[3,0],[0,7]][[3, 0], [0, 7]] (rows)?
    P=IP = I and D=[[3,0],[0,7]]D = [[3, 0], [0, 7]]. A diagonal matrix is already diagonalized by the standard basis.
  12. What does it mean for a square matrix AA to be orthogonally diagonalizable?
    A=PDPTA = PDP^T for some orthogonal matrix PP (so P−1=PTP^{-1} = P^T) and diagonal matrix DD.

Study these flashcards