Given a sequence of number $\{ a_{n}\} $, define an operator on $l^2$ as follows: $$A\left( x_{1},x_{2},...,x_{n},...\right) =\left( a_{1}x_{1},a_{2}x_{2},...,a_{n}x_{n},...\right) $$ If $A$ is a bounded operator, determine the spectral point type of $A$.
My attempt:
It's easy to see that $\{ a_{n}\} $ is bounded if $A$ is a bounded operator, but I think this condition is not enough to determine the whole spectrum, it depends on the property of $\{ a_{n}\} $, for example, $a_{n}=\frac{1}{n} $, we can determine the whole spectrum.
I claim first that $\{a_j:j\in\mathbb{N}\}$ is the set of eigenvalues of $A$. Indeed it is clear that each $a_j$ is an eigenvalue, as if $e_j\in\ell^2$ is $1$ in the $j$'th position, and $0$ everywhere else, then
$$Ae_j=a_je_j.$$
To see that any eigenvalue is in the set, suppose $\lambda$ is an eigenvalue of $A$ with eigenvector $x=\{x_j\}_{j\in\mathbb{N}}$, then
$$Ax=\{a_jx_j\}_{j\in\mathbb{N}}=\{\lambda x_j\}_{j\in\mathbb{N}},$$
so $\lambda x_j=a_j x_j$ for all $j\in\mathbb{N}$, and so in particular, as $x\neq0$, $\lambda=a_j$ for some $j\in\mathbb{N}$. This proves the claim.
We show now that $\sigma(A)=\overline{\{a_j:j\in\mathbb{N}\}}$. It is clear that
$$\overline{\{a_j:j\in\mathbb{N}\}}\subseteq\sigma(A)$$
as $\sigma(A)$ is a closed set containing all eigenvalues of $A$. We show now that
$$\mathbb{C}\setminus\overline{\{a_j:j\in\mathbb{N}\}}\subseteq\rho(A).$$
Let $\lambda\in\mathbb{C}\setminus\overline{\{a_j:j\in\mathbb{N}\}}$, and find $\varepsilon>0$ such that $\lvert\lambda-a_j\rvert\geq\varepsilon$ for all $j\in\mathbb{N}$. Then
$$(\lambda I-A)\{x_j\}_{j\in\mathbb{N}}=\{(\lambda-a_j)x_j\}_{j\in\mathbb{N}},$$
so we have the inverse
$$(\lambda I-A)^{-1}\{x_j\}_{j\in\mathbb{N}}=\left\{\frac{1}{\lambda-a_j}x_j\right\}_{j\in\mathbb{N}},$$
which is a well defined operator $\ell^2\to\ell^2$ as $\frac{1}{\lambda-a_j}\leq\frac{1}{\varepsilon}$ for all $j\in\mathbb{N}$, and so $\lambda\in\rho(A)$. It follows that
$$\sigma(A)=\overline{\{a_j:j\in\mathbb{N}\}}.$$