I have an operator $$T\in \mathcal L(\mathbb{F}^n)$$ defined as $$\ T(x_1, x_2, \ldots, x_n) = (x_1, 2x_2, 3x_2,\ldots, n x_n)$$ $\mathbb{F}$ is either $\mathbb{R}$ or $\mathbb{C}$.

This is from Linear Algebra Done Right, and I'm tasked with finding a) the eigenvectors and values, and b) all the invariant subspaces...

a) I'm pretty sure we have eigenvalue-eigenvector pairs $$1, (1, 0,\ldots,0)\\2,(0,1,\ldots,0)\\\vdots\\ n, (0,0,\ldots,n)$$ found by solving the system of equations for $T(x_1,x_2,\ldots,x_n)=\lambda(x_1,x_2,\ldots,x_n)$.

b) is what I'm having trouble with. I'm inclined to believe it only has linear combinations of the eigenvectors and the obvious ones (like $\{0\}$) as subspaces.

I think that if you have a subspace $W$ and $k$ vectors $$W = \{a_0 v_0 + \ldots + a_k v_k | a_i \in \mathbb{F}, v_i \in W,i=1,\ldots,k\}$$ then its only invariant if there are $v_0,\ldots,v_k$ themselves invariant under $T$?

Because then $$\begin{align}Tw &= a_0 Tv_0 + \ldots + a_k Tv_k\\Tw &= a_0 \lambda_0 v_0+\ldots+a_k \lambda_k v_k\end{align}$$ If it's not, we have a sum involving a vector not in $W$.

Am I on the right track?


Since all eigenvalues are different then any set of eigenvectors define an invariant subspace. So, there are $2^n$ such invariant subspaces.

  • $\begingroup$ How would I know these are all there is though? I am expected to solve this with the bare minimum: distinct eigenvectors are linearly independent, there at most as many values as the dimension of the space (for finite spaces).. but I'm probably not seeing something obvious. $\endgroup$ – Skurmedel Dec 4 '16 at 9:17

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.