I A true or false statement concerning condition number of a matrix

  • I
  • Thread starter Thread starter psie
  • Start date Start date
  • Tags Tags
    Linear algebra
psie
Messages
315
Reaction score
40
TL;DR Summary
I'm reading Linear Algebra by Friedberg et al. They have a section on conditioning and the Rayleigh quotient. At the end, there's an exercise with a couple of true or false statement, and I'm stuck with two of them, since the answer at the back of the book seems contradictory to me.
Conditioning doesn't refer to hair conditioners in this case, but what happens to the solution of a system of linear equations ##Ax=b## when we change say the vector ##b## slightly to ##b'##. If the relative change in ##b##, that is ##\frac{\|b-b'\|}{\|b\|}##, is close to the relative change in ##x##, i.e. ##\frac{\|x-x'\|}{\|x\|}##, the system is called well-conditioned; otherwise it is ill-conditioned. If ##A## is invertible and ##b\neq 0##, then we have $$\frac1{\operatorname{cond}(A)}\frac{\|b-b'\|}{\|b\|}\leq \frac{\|x-x'\|}{\|x\|}\leq \operatorname{cond}(A)\frac{\|b-b'\|}{\|b\|},$$where ##\operatorname{cond}(A)=\|A\| \|A^{-1}\|## is the condition number of ##A##. This number is ##\geq 1## and depends on the matrix norm (the preceding inequality, however, holds for any norm satisfying ##\|Ax\|\leq\|A\| \|x\|## I think). I know at least for the Euclidean norm, that ##\|A\|## equals the square root of the largest eigenvalue of ##A^\ast A##, where ##A^\ast## is the conjugate transpose. I've heard that computing ##\operatorname{cond}(A)## is a tricky business in general.

Anyway, now I'm faced with the following two statements:

(a) If ##Ax=b## is well-conditioned, then ##\operatorname{cond}(A)## is small.

(b) If ##\operatorname{cond}(A)## is large, then ##Ax=b## is ill-conditioned.

My book says (a) is false. I agree, since if it is well-conditioned, the relative change in ##b## and ##x## has ratio near ##1## or at least ##1## plus some remainder that isn't very large in absolute value (my book gives a somewhat loose definition of well- and ill-conditioned systems, so I'm not very familiar what counts as ill-conditioned in terms of numbers). By the displayed inequality above, if we divide ##\frac{\|b-b'\|}{\|b\|}##, then we obtain $$\frac1{\operatorname{cond}(A)}\leq 1+h \leq \operatorname{cond}(A)$$where ##|h|\geq0## is small. I don't see that the preceding inequality implies that the condition number has to be small, so I conclude the statement is false.

Now notice, I think (b) is just the contrapositive of (a). However, my book claims it is true. It has been wrong before, but here I'm not so sure anymore.
 
Physics news on Phys.org
psie said:
Now notice, I think (b) is just the contrapositive of (a). However, my book claims it is true. It has been wrong before, but here I'm not so sure anymore.
Actually, according to what I can find online, your reference labels both (a) and (b) as False. The only True statement is:
(c) If cond(A) is small, then Ax = b is well-conditioned.
These truth-values follow from the statement on pg. 470:
1748549176580.webp
 
renormalize said:
Actually, according to what I can find online, your reference labels both (a) and (b) as False.
Ah ok, then I have an earlier printing of the book. It actually says (b) is true in my book. Thank you!
 
I asked online questions about Proposition 2.1.1: The answer I got is the following: I have some questions about the answer I got. When the person answering says: ##1.## Is the map ##\mathfrak{q}\mapsto \mathfrak{q} A _\mathfrak{p}## from ##A\setminus \mathfrak{p}\to A_\mathfrak{p}##? But I don't understand what the author meant for the rest of the sentence in mathematical notation: ##2.## In the next statement where the author says: How is ##A\to...
The following are taken from the two sources, 1) from this online page and the book An Introduction to Module Theory by: Ibrahim Assem, Flavio U. Coelho. In the Abelian Categories chapter in the module theory text on page 157, right after presenting IV.2.21 Definition, the authors states "Image and coimage may or may not exist, but if they do, then they are unique up to isomorphism (because so are kernels and cokernels). Also in the reference url page above, the authors present two...
When decomposing a representation ##\rho## of a finite group ##G## into irreducible representations, we can find the number of times the representation contains a particular irrep ##\rho_0## through the character inner product $$ \langle \chi, \chi_0\rangle = \frac{1}{|G|} \sum_{g\in G} \chi(g) \chi_0(g)^*$$ where ##\chi## and ##\chi_0## are the characters of ##\rho## and ##\rho_0##, respectively. Since all group elements in the same conjugacy class have the same characters, this may be...
Back
Top