Display Accessibility Tools

Accessibility Tools

Grayscale

Highlight Links

Change Contrast

Increase Text Size

Increase Letter Spacing

Readability Bar

Dyslexia Friendly Font

Increase Cursor Size

Numerical Analysis - May 2024

1.

(15 points) Let A , B ∈ ℂ m × m be arbitrary matrices. Show that

∥ A B ∥ F ≤ ∥ A ∥ 2 ∥ B ∥ F

where ∥ . ∥ 2 and ∥ . ∥ F denote the 2-norm and Frobenius norm, respectively.

2.

(15 points) Let A ∈ ℂ m × n with m ≥ n . Show that A * A is nonsingular if and only if A has full rank.

3.

(15 points) Answer the following questions about properties of matrices.

  • Let A , B ∈ ℂ n × n , with A nonsingular and ∥ A − 1 ∥ ∥ B ∥ = q < 1 . Show that A + B is nonsingular, i.e., invertible.
  • Let A ∈ ℂ n × n be skew-Hermitian, i.e., A * = − A . Show that the matrix I + A is nonsingular and that ( I + A ) − 1 ( I − A ) is unitary.

4.

(15 points) Let P ∈ ℂ m × m be a projector. Show P is orthogonal if and only if P = P * .

5.

(10 points) Suppose A is an invertible 202 × 202 matrix with ∥ A ∥ 2 = 100 and ∥ A ∥ F = 101 . Give the sharpest possible lower bound on the 2-norm condition number κ ( A ) .

6.

(15 points) Prove Gerschgorin’s disk theorem: Let A ∈ ℂ m × m with entries { a i j } and let r i = ∑ j ≠ i | a i j | . Let D i be the closed disk in ℂ with center a i i and radius r i . If λ is an eigenvalue of A, then λ ∈ ∪ i D i ; in other words, λ lies within at least one of the disks D i .

7.

(15 points) Suppose the Arnoldi iteration is executed for a particular m × m matrix A and vector b until at some step n, an entry h n + 1 , n = 0 is encountered.

  • Show how the equality A Q n = Q n + 1 H̃ n can be simplified in this case. Recall that Q n is an m × n matrix whose columns are the first n columns of Q in A Q = Q H , where H is a Hessenberg matrix. In addition, H̃ n is the ( n + 1 ) × n upper-left section of H.
  • Show that K n is an invariant subspace of A, i.e. A K n ⊆ K n , where K n is the n th Krylov subspace, which is the span of { b , A b , … A n − 1 b } ?