Hilbert-Schmidt integral operator is compact.

In what follows, we will prove that every Hilbert-Schmidt integral operator is compact.
Formally, let $k\in L_2([0,1]\times[0,1])$ and $T:L_2[0,1]\to L_2[0,1]$ be defined by
$$Tf(x):=\int_0^1k(x,y)f(y)dy.$$
Then $T$ is compact operator.

Although the proof may seem a bit long at first sight, it is proceeds naturally, utilizing approximation of $k$ with continuous functions.

$\textbf{Proof}$. We approximate $k$ with a sequence of continuous functions, then prove that each of the corresponding operators has image contained in $C[0,1]$, then show that each of these operators is compact when $C[0,1]$ is equipped with the $\sup$ norm. This will be done via the Arzela-Ascoli theorem. Finally show that these converge uniformly (in the operator norm) to the original operator, and since they are compact, the original must be as well.
The Cauchy-Bunyakovsky-Schwartz inequality is used throughout the proof.

Thus we start by choosing a sequence $(k_n)_{n\ge 1}$ of continuous functions converging to $k$ in the sense of $L_2$ norm, i.e.
$$\lim_{n\to\infty}\int_0^1(k_n(x)-k(x))^2\text{d}x=0.$$
Such a sequence could be constructed using the Fourier decomposition of $k$ into sines for example.
Then for any $n\ge 1$ define $$T_nf(x):=\int_0^1k_n(x,y)f(y)dy.$$
Fix $m\ge 1$ and for brevity denote $S=T_m$.
First we show that $\text{Im} S\subseteq C[0,1]$. Let $f\in L_2[0,1]$. We aim to show that $Sf$ is a continuous function. Indeed, let $x\in [0,1]$ and $(x_n)_{n\ge 1}$ be a sequence in $[0,1]$ converging to $x$. Then, if we denote $g_n(y)=k_m(x_n,y)$, we have
$$\lim_{n\to\infty}g_n(y)=k_m(x,y)$$
Moreover, if $\displaystyle C:=\sup_{x\in[0,1]}|k_m(x,y)|$ (finite, since $k_m$ is continuous), then $|g_n(y)|=|k_m(x_n,y)|\le C$. Thus Dominated convergence theorem is applicable, hence
$$\lim_{n\to\infty}Sf(x_n)=\lim_{n\to\infty}\int_0^1g_n(y)f(y)\text{d}y=\int_0^1k_m(x,y)f(y)\text{d}y=Sf(x).$$
So $S:L_2[0,1]\to C[0,1]$.
Next, equip $C[0,1]$ with the $\sup$ norm. We will prove $S$ is compact. To this end, we take a sequence $(f_n)_{n\ge 1}\subset L_2[0,1]$, such that $\|f_n\|_2\le 1$ for all $n$, and we are going to prove that $(Sf_n)_{n\ge 1}$ has a convergent subsequence in $C[0,1]$. The criterion telling us whether this is true is the Arzela-Ascoli theorem.
From the Cauchy-Bunyakovsky-Schwartz inequality we obtain $$|Sf_n(x)|=\left|\int_0^1 k_m(x,y)f_n(y)\text{d}y\right|\le \sqrt{\int_0^1k_m(x,y)^2\text{d}y\int_0^1f_n(y)^2\text{d}y}\le C$$
(for $\displaystyle C:=\sup_{x\in[0,1]}|k_m(x,y)|$) and thus the sequence $(Sf_n)_{n\ge 1}$ is uniformly bounded.
Moreover, let $\varepsilon>0$. Then there is some $\delta>0$ such that whenever $|x_1-x_2|< \delta$, then $|k_m(x_1,y)-k_m(x_2,y)|<\varepsilon$ for all $y\in[0,1]$ (uniform continuity of continuous functions).Thus for any $n$ and any $x_1,x_2$, such that $|x_1-x_2|<\delta$ we have
$$|Sf_n(x_1)-Sf_n(x_2)|=\left|\int_0^1(k_m(x_1,y)-k_m(x_2,y))f_n(y)\text{d}y\right|$$
$$\le \sqrt{\int_0^1(k_m(x_1,y)-k_m(x_2,y))^2\text{d}y \int_0^1 f_n(y)^2\text{d}y}\le \varepsilon$$
This shows that the sequence $(Sf_n)_{n\ge 1}$ is equicontinuous. Thus it contains convergent subsequence.
This concludes the proof that $(T_m)_{m\ge 1}$ is a sequence of compact operators (acting from $L_2[0,1]$ to $C[0,1]$). Moreover, for any $f\in C[0,1]$, $\|f\|_{\infty}\ge \|f\|_2$, which means that the $L_2$ topology on $C[0,1]$ is weaker than the uniform topology. Hence the operators $T_m$, $m\ge 1$, are compact, even if we consider them from $L_2[0,1]$ to $L_2[0,1]$.
It remains to observe that $T_m$ converges to $T$ in the operator norm.
Let $f\in L_2[0,1]$. Then
$$\|T_mf-Tf\|_2^2=\int_0^1\left(\int_0^1(k(x,y)-k_m(x,y))f(y)\text{d}y\right)^2\text{d}x\le$$
$$\int_0^1\int_0^1(k(x,y)-k_m(x,y))^2\text{d}y\int_0^1f(y)^2\text{d}y\text{d}x=\int_0^1\int_0^1(k(x,y)-k_m(x,y))^2\text{d}y\text{d}x\int_0^1f(y)^2\text{d}y=$$
$$\|k_m-k\|_2^2\|f\|_2^2.$$
This shows that
$$\|T_m-T\|\le \|k_m-k\|_2,$$
and since $(k_m)_{m\ge 1}$ was taken to converge to $k$ in $L_2$, then $T_m$ converges to $T$.
The proof is finished, by saying, that the space of compact operators is closed, and as $T_m$ are compact, so is $T$.

Comments

Post a Comment

Popular posts from this blog

Basel problem type sum, proposed by prof. Skordev

Evaluate $$\sum_{n=1}^{\infty}\frac{1}{2^n n^2}.$$ Proof. The function $$S(x):=\sum_{n=1}^{\infty}\frac{x^n}{ n^2}$$ is well defined for $x\in[-1,1]$ and thus we need to find $S(1/2)$. Using differentiation, we observe that $$S(x)=-\int_0^x\frac{\log(1-t)}{t}\text{d}t.$$ If we make change of variables in the latter integral $t\to 1-t$ we obtain $$S(x)=-\int_{1-x}^1\frac{\log t}{1-t}\text{d}t. \ \ \ \ \ (1)$$ On the other hand, using integration by parts, we obtain $$S(x)=-\int_0^x\log(1-t)\text{d}\log t=-\log t\log(1-t)\Big|_{t=0}^x-\int_0^x\frac{\log t}{1-t}\text{d}t.$$ Notice that $\displaystyle \lim_{t\to 0}\log t \log(1-t)=0$ ($\log (1-t)\sim t$ and then L'Hopital), hence $$S(x)=-\int_0^x\log(1-t)\text{d}\log t=-\log x\log(1-x)-\int_0^x\frac{\log t}{1-t}\text{d}t. \ \ \ \ \ (2)$$ Plugging $x=1/2$ and equating $(1)$ to $(2)$ we obtain $$\int_{1/2}^1\frac{\log(1-t)}{t}\text{d} t-\int_{0}^{1/2}\frac{\log(1-t)}{t}\text{d} t= \log\left(\frac{1}{2}\right)^2.$$ On the ot
Read more

Modification on a sequence from VJIMC

The first two limits of the following problem were proposed at VJIMC, 2005, Category I. The exact value of the last limit was proposed by a user at https://artofproblemsolving.com , where you can also see his solution along other lines.  Let $(x_n)_{n\ge 2}$ be a sequence of real numbers, such that $x_2>0$ and for every $n\ge 2$ holds  $$x_{n+1}=-1+\sqrt[n]{1+nx_n}.$$ Prove consecutively that  $$1)\ \lim_{n\to\infty}x_n=0,\  \ \ 2)\ \lim_{n\to\infty}nx_n=0,\  \ \ 3)\ \lim_{n\to\infty}n^2x_n=4.$$ $\textbf{Proof.}$  $\textbf{1)}$ Clearly all the elements of the sequence are positive. The inequality $-1+\sqrt[n]{1+nx_n}<x_n$ is equivalent to $(1+nx_n)<(1+x_n)^n$, which is seen to be true after expanding, since all the summands on the right are positive. This shows that the sequence is strictly decreasing. Hence $$0<x_{n+1}=-1+\sqrt[n]{1+nx_n}\le -1+\sqrt[n]{1+nx_2},$$ and since the right hand side clearly tends to $0$ we obtain $\displaystyle\lim_{n\to\infty}x_n=0.$ $\textbf{
Read more

A problem proposed by prof. Babev

Consider the sequence $$a_n=\int_{n}^{n+1}\frac{\sin^2(\pi t)}{t}\text{d}t.$$ It could be proven that $$\lim_{n\to \infty}n a_n=\frac{1}{2}$$  Now we are going to prove that $$\lim_{n\to\infty}n\left(na_n-\frac{1}{2}\right)=-\frac{1}{4} $$ and show how to derive all such limits. $\textbf{Proof.}$ Since $\displaystyle\int_{n}^{n+1}\sin^2(\pi t)\text{d}{t}=\frac{1}{2}$ (for any $n\in \mathbb{N}$) we can rewrite the limit in question as $$\lim_{n\to \infty}n\left(n\left(\int_{n}^{n+1}\frac{\sin^2(\pi t)}{t}\text{d}t-\int_{n}^{n+1}\frac{\sin^2(\pi t)}{n}\text{d}t\right)\right)=\lim_{n\to \infty}n^2\left(\int_{n}^{n+1}\sin^2(\pi t)\left(\frac{1}{t}-\frac{1}{n}\right)\text{d}t\right)$$ $$=\lim_{n\to \infty}n\left(\int_{n}^{n+1}\sin^2(\pi t)\frac{n-t}{t}\text{d}t\right)$$ Changing the variables $t\to n+s$ and using periodicity of sine, the last simplifies to $$ \lim_{n\to \infty}n\int_{0}^{1}\sin^2(\pi s)\frac{-s}{n+s}\text{d}s$$ Since for $s\in[0,1]$, $\displaystyle \frac
Read more