Yuze Zhao bio photo

Email

My CV

Github

Functional Analysis 11.13

13. Use the Uniform Boundedness Principle to prove the Hellinger–Toeplitz Theorem.

Construct a functional $f_y$ such that

\[f_y(x)=(Ax,y)=(x,Ay).\]

In fact, they satisfy the norm relation $f_y = \Vert Ay \Vert$. We first prove this relation:

\[\Vert f_y\Vert=\sup_{\Vert x \Vert =1}|(x,Ay)|\leq \Vert x\Vert\cdot\Vert Ay \Vert=\Vert Ay \Vert.\]

Moreover, we also have

\[f_y(Ay)=(Ay,Ay)=\Vert Ay\Vert^2,\]

which implies that $\Vert f_y\Vert\geq \Vert Ay \Vert$. Thus, in fact, $\Vert f_y\Vert=\Vert Ay \Vert$. Now we prove the theorem. Since it is defined everywhere, we obtain

\[\sup_{\Vert y\Vert= 1}\Vert f_y(x)\Vert=\sup_{\Vert y\Vert= 1}|(Ax,y)|\leq\Vert Ax\Vert<\infty.\]

We regard ${f_y}$ as a family of bounded functionals indexed by $y$. By the Uniform Boundedness Principle together with the definition of the operator norm in an inner product space,

\[\Vert A\Vert=\sup_{||y||=1}\Vert Ay\Vert=\sup_{\Vert y\Vert= 1}\Vert f_y\Vert<\infty.\]

This shows that $A$ is bounded.


14. Suppose $A,B$ are linear operators defined everywhere on a Hilbert space $H$, and satisfy

\((Ax,y)=(x,By),\quad \forall x,y\in H.\)

Prove: $A,B$ are bounded and $B=A^*$.


The proof idea is similar to the previous problem.

\[\Vert f_y\Vert=\sup_{\Vert x \Vert =1}|(x,By)|\leq \Vert x\Vert\cdot\Vert By \Vert=\Vert By \Vert,\]

and

\[f_y(By)=(By,By)=\Vert By\Vert^2.\]

Hence we also have $\Vert f_y\Vert=\Vert By \Vert$. Moreover,

\[\sup_{\Vert y\Vert= 1}\Vert f_y(x)\Vert=\sup_{\Vert y\Vert= 1}|(Ax,y)|\leq\Vert Ax\Vert<\infty.\]

Exactly as in Problem 13, by the Uniform Boundedness Principle, $B$ is bounded.

Since

\[(Ax,y)=(x,By),\]

it follows that $A$ is also bounded. By the definition of the adjoint operator, we have $B=A^*$.


15. Suppose ${x_n}_{n=1}^\infty$ is a sequence in a Banach space $X$, and for every $f\in X’$,

\[\sum^{\infty}_{n=1}|f(x_n)|<\infty,\]

then there exists a constant $\mu>0$ such that for all $f\in X’$,

\[\sum^\infty_{n=1}|f(x_n)|\leq \mu\ \|f\|.\]

First, let us clarify that $X’$ denotes the dual space, i.e. the space of all bounded linear functionals on $X$.

Now, construct a sequence of operators $T_n\in \mathcal{L}(X’,l^1)$, which are bounded linear operators from the Banach space $X’$ to the normed space $l^1$, defined by

\[T_k(f)=\sum^k_{n=1}|f(x_n)|,\qquad k\in \mathbb{N^*}.\]

By assumption,

\[\sup_{k\in \mathbb{N^*}}|T_k(f)|=\sum^{\infty}_{n=1}|f(x_n)|<\infty.\]

By the Uniform Boundedness Principle,

\[\sup_n \|T_n\|\leq \mu <\infty.\]

Therefore, we have

\[\sum^\infty_{n=1}|f(x_n)|=\lim_{n\to\infty}T_n(f)\leq \|T_n\|\ \|f\|\leq \mu \|f\|.\]

Hence the claim is proved.