0
$\begingroup$

In Kuo's book Gaussian measures in Banach spaces, he says the space $\mathbb{R}^{[0,1]}$ the set of all real valued functions on $[0,1]$ with $x(0)=0$ is a separable Banach space under the norm $$\|x\|=\sup_{t}|x(t)|$$

I know that $\mathbb{R}^{[0,1]}$ is separable by The Engelking-Karlowicz theorem. But how can I show the space given above is open in it? Is this space really Banach?

The problem is, it is possible to define the Wiener process on $\mathbb{R}^{[0,1]}$ with a cylinder sigma algebra and show there exists a version on $C[0,1]$. I am just confused about the method used in the book.

enter image description here

$\endgroup$
2
  • 1
    $\begingroup$ I don't see why $\sup_{t\in[0,1]} \lvert x(t)\rvert$ should be finite for a real-valued function $x$ (let even $x(0)=0$, for all it matters). Typically, $C[0,1]$ denotes the set of continuous real-valued functions, so perhaps there is a typo in the book. $\endgroup$ – Gae. S. Sep 11 '19 at 23:40
  • 1
    $\begingroup$ Exactly why I'm asking. He doesn't say continuous $\endgroup$ – user515599 Sep 11 '19 at 23:40
2
$\begingroup$

The topological space $\Bbb R^{[0,1]}$ (which is also homeomorphic to the subspace of the functions such that $x(0)=0$) is not first-countable, therefore it is not metrisable. There must be a typo in the book, and $C[0,1]$ should stand for a space of continuous functions.

$\endgroup$
0

Your Answer

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