Exercise 2.1.5

Answers

For b = (1,0,0), Let’s run the Arnoldi iteration with A to produce an orthonormal basis q1,q2,q3: A = [ 2 1 0 1 2 1 0 1 2 ]

  • k = 1, Let q1 = b |b| = (1,0,0).
  • v = Aq1 = (2,1,0)

    • j = 1: h11 = q1Tv = 2, v = vh11q1 = [ 0 1 0 ]
  • h2,1 = 1
  • q2 = v h2,1 = [ 0 1 0 ]
  • k = 2, we have v = Aq2 = [ 1 2 1 ]

    • j = 1: h12 = q1Tv = 1, $v=v-h_{12}q_1 =

      [ 0 2 1 ]

      $

    • j = 2: h22 = q2Tv = 2, $v=v-h_{22}q_2 =

      [0 0 1 ]

      $

  • h3,2 = 1
  • $q_3 =

    [0 0 1 ]

    $

So from the calculations, we have H = [ 2 1 1 2 0 1 ]. We can prove that AQ2 = Q3H

User profile picture
2020-03-20 00:00
Comments