# Thread: linear functionals, closed sets, convergence

1. ## linear functionals, closed sets, convergence

This is the last round of questions, I promise.

(1) Identify the space of all continuous linear functionals on the space $c_0$.

(2) $X=(C([0, 2], \mathbb{R}), || . ||_2)$,
$Y=\{f \in C([0, 2], \mathbb{R}: f(1-x)=f(1+x), x \in [0,1]\}$
Is Y closed vector space in X? What is the complement of Y?

(3) Prove that the series
$\sum_{k=1}^{\infty}x_k b_k$ converge for every vector
$x=(x_k) \in l^p, 1 if and only if $b=(b_k) \in l^q, 1/p +1/q=1$.

These are the last ones I don't know what to do with, and if anyone has the time and is willing to help, I would be really grateful.

2. Originally Posted by marianne
(1) Identify the space of all continuous linear functionals on the space $c_0$.
The dual space of $c_0$ is (isometrically isomorphic to) $l^1$. If $x=(x_n)\in c_0$ and $y=(y_n)\in l^1$ then the value of the functional y at the point x is $\sum x_ny_n$. This is a very standard result which you should be able to find in any relevant textbook.

Originally Posted by marianne
(2) $X=(C([0, 2], \mathbb{R}), || . ||_2)$,
$Y=\{f \in C([0, 2], \mathbb{R}): f(1-x)=f(1+x), x \in [0,1]\}$
Is Y closed vector space in X? What is the complement of Y?
This is more interesting. Take the second question first. If g(x) is in the orthogonal complement of Y, then $\int_0^2f(x)g(x)\,dx=0$ for all f in Y. But $\int_0^2f(x)g(x)\,dx = \int_0^1f(x)g(x)\,dx + \int_1^2f(x)g(x)\,dx$. If you change variables in the two integrals, putting y=1-x in the first and y=1+x in the second, then you find that $\int_0^1f(1-y)\bigl(g(1-y)+g(1+y)\bigr)dy=0$ (remembering that f(1+y)=f(1-y)). This holds for all continuous functions on [0,1], so you conclude that g(1+y)=–g(1-y) for all y in [0,1].

Thus the orthogonal complement of Y is $Z=\{g \in C([0, 2], \mathbb{R}): g(1-x)=-g(1+x), x \in [0,1]\}.$ If you repeat the same argument on the subspace Z, you will find that its orthogonal complement is Y. Therefore Y is equal to its second orthogonal complement, which is closed in X. Hence Y is closed.

Originally Posted by marianne
(3) Prove that the series
$\sum_{k=1}^{\infty}x_k b_k$ converge for every vector
$x=(x_k) \in l^p, 1 if and only if $b=(b_k) \in l^q, 1/p +1/q=1$.
This is another standard piece of bookwork. Look it up in any good textbook. The proof uses Hölder's inequality.