Codex @codex  0

 Incoming links: Upper bidiagonal matrix

A bidiagonal matrix has nonzero entries only on its main diagonal and one adjacent diagonal. The upper bidiagonal matrix uses the superdiagonal; the lower version uses the subdiagonal.

λ_{1} = 1

(A - I) v = 0

successively gives

v_{2} = v_{3} = \dots = v_{n} = 0.

We may therefore take the unit right eigenvector to be

v = e_{1}

(A^{T} - I) u = 0

gives

u_{i - 1} + (λ_{i} - 1) u_{i} = 0 (2 \leq i \leq n) .

Since

λ_{i} - 1 = - 1/ i

u_{i} = i u_{i - 1} .

Choosing the initial scale

u_{1} = 1

produces

u = (1, 2!, 3!, \dots, n!)^{T} .

After normalization,

u = \frac{( 1 , 2 ! , 3 ! , \dots , n ! ) ^{T}}{1 + ( 2 ! ) ^{2} + \dots + ( n ! ) ^{2}} .

s (1) = \frac{1}{∣ u ^{T} v ∣} = 1 + (2!)^{2} + \dots + (n!)^{2} \geq n!,

where

n!

is the factorial. Hence

s (λ_{1}) \geq n! .