Sign up ×
MathOverflow is a question and answer site for professional mathematicians. It's 100% free, no registration required.

Let $B$ be a Banach space. Let $\{Y_{n}\}$ be a sequence of $B$ valued random variables.


  1. $P(\{Y_{n}\} \mbox{is bounded}) = 1$,

  2. fo every $\epsilon>0$, there exists a finite dimensional subspace $F$ such that $P(\limsup_{n} q_{F}(Y_{n})\leq \epsilon) = 1$.

Then show that $P(\{Y_{n}\} \mbox{ is relatively compact}) = 1$. Where $q_{F}(x) = d(x,F)$ is the distance of $x$ from $F$.

share|cite|improve this question
its not HW....I came across this in the proof of Law of iterated logarithms for Banach Space valued RVS by Kuelbs....I know this would be easy to prove..... – user24367 Feb 18 '13 at 0:28
Which part can you not prove? What have you tried so far? See – Yemon Choi Feb 18 '13 at 2:03
Also, check your statement ... fix the conclusion, perhaps... – Gerald Edgar Feb 18 '13 at 14:26

1 Answer 1

up vote 0 down vote accepted

For $\omega\in\Omega$, define $S(\omega):=\(Y_n(\omega),n\in\mathbb N\)$. We have to show that $P(\omega\mid S(\omega)\mbox{ is relatively compact})=1$.

Considering $\varepsilon=1/j$, we can see that there is $\Omega'$ of probability $1$ such that for each $\omega\in \Omega'$,

  1. $\sup_n\lVert Y_n(\omega)\rVert<\infty$, and
  2. for each integer $j$, $\limsup_{n\to +\infty}d(Y_n(\omega),F_j)\leqslant \frac 1j$, where $F_j$ corresponds to the subspace for $\varepsilon=\frac 1j$.

We say that a set $S\subset B$ is flately concentrated if for each $ \varepsilon>0$, we can find a finite dimensional subspace $F$ of $B$ such that for all $x\in S$, $d(x,F)<\varepsilon$.

It's linked to relative compactness (using completeness) by the following:

A subset $S$ of the Banach space $B$ has a compact closure if and only $S$ is bounded and flately concentrated .

To conclude here, we have to check that $S(\omega)$ is flately concentrated for each $\omega\in\Omega'$. Argue by contradiction (the problem will come from infinitely many $n$).

share|cite|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.