I figure it has something to do with minimal polynomials and the Jordan canonical form, I just can't piece it together. I would really appreciate a hint on how to start it.
Asked
Active
Viewed 45 times
2 Answers
2
If the range of $X^k$ is equal to the range of $X^{k+1}$ then the range of $X^{k+i}$ is equal to the range of $X^k$ for all positive $k$.
It follows the ranks of $X,X^2,\dots,X^b$ are strictly decreasing, and since the rank of $X$ is at most $n$ we have $b$ is at most $n$.
-
2Excellent! I thought this was much deeper, using Cayley-Hamilton to bound the degree of the minimal polynomial... – David C. Ullrich May 09 '21 at 15:01
-
1I think Cayley Hamilton uses something similar in one of the parts of the proof with $(X-\lambda I)^k$ on some of the subspaces but I don't remember very well. – Asinomás May 09 '21 at 15:03
0
Your hunch is correct: The minimal polynomial must be divide $t^m$ and have degree at most $n$.
lhf
- 216,483