Proving/disproving $\|A\|^{-1} = \|A^{-1}\|$ for non-singular $A$

75 Views Asked by At

Cannot find this question, though I find it hard to believe that it hasn't been answered somewhere. All I really have here are the requirements for something to be considered a matrix norm, and that my A must be non-singular.

I got that $||A||\cdot ||A^{-1}|| \le ||AA^{-1}|| = ||1||$ from the triangle inequality, so if I could drop that pesky $\le$ operator and replace it with proper equality, I would know that $||A^{-1}||$ was in fact the inverse of $||A||$, a.k.a.$||A^{-1}|| = ||A||^{-1}$. But I can't think of how to show equality, or if I even can.

I also may be making bad assumptions in that $||1|| = 1$, because the identity matrix may not necessarily map to 1? Though I can't see why it wouldn't. My linear is very shaky, hopefully I'm going in the right direction.

1

There are 1 best solutions below

2
On BEST ANSWER

It is rarely true that the norm of the inverse is the inverse of the norm. E.g., for $A$ the diagonal two-by-two diagonal matrix with diagonal entries $1$ and $2$, the norm is $2$. The inverse has diagonal entries $1$ and $1/2$, and has norm $1$.

EDIT: here $\|A\|=\sup_{|x|\le 1} |Ax|$, the usual operator norm.