Proof of Stationary Values for Ritz Values in Lanczos Iteration

  • Level: Graduate 
  • Thread starter Thread starter redlegend
  • Start date Start date
Join the discussion
Registration is free. Start your own thread to ask a follow-up.
1 reply · 3K views
redlegend
Messages
1
Reaction score
0
I was wondering whether you have the proof to the following lemma:

"The Ritz values at step n of the Lanczos iteration are the stationary values of the Rayleigh quotient r(x)=(xT A x)/(xT x) if x is restricted to Kn "
 
Physics news on Phys.org
Did you ever find the proof, because I have the same exact question?