Filtration's Lp Spaces are Dense in Limit Filtration's Lp Space
Theorem
Let $\struct {X, \Sigma, \mu}$ be a measure space.
Let $\sequence {\FF_n}_{n \mathop \in \N}$ be a filtration of $\Sigma$.
Let $\FF_\infty$ be the limit of $\sequence {\FF_n}_{n \mathop \in \N}$.
Let $p \ge 1$.
Let $\map {L^p} {\cdot}$ denote the $L^p$ spaces.
Then $\ds \bigcup_{n \mathop \ge 0} \map {L^p} {X, \FF_n, \mu}$ is a dense subset of $\map {L^p} {X, \FF_\infty, \mu}$.
Proof
First, it is a subset, since
- $\forall n \ge 0 : \map {L^p} {X, \FF_n, \mu} \subseteq \map {L^p} {X, \FF_\infty, \mu}$
in view of $\FF_n \subseteq \FF_\infty$.
In the following, we show its density.
Let:
- $\ds \AA_0 := \bigcup_{n \mathop \ge 0} \FF_n$
Then $\AA_0$ is an algebra.
By Sigma-Algebra extended from Algebra by Measure:
- $\ds \BB := \bigcap_{\epsilon \mathop > 0} \bigcup_{A \mathop \in \AA_0} \set {B \in \Sigma : \map \mu {A \symdif B} \le \epsilon}$
is a $\sigma$-algebra.
In particular:
- $(1):\quad \FF_\infty = \map \sigma {\AA_0} \subseteq \BB$
Let $f \in \map {L^p} {X, \FF_\infty, \mu}$.
Let $\epsilon > 0$.
By Space of Simple P-Integrable Functions is Everywhere Dense in Lebesgue Space, there exist:
- $n \in \Z_{>0}$
- $\alpha_1, \ldots, \alpha_n \in \R_{\ne 0}$
- $B_1, \ldots, B_n \in \FF_\infty$
such that:
- $\ds (2): \quad \norm {f - \sum_{k \mathop = 1}^n \alpha_k \chi_{B_k} }_p \le \epsilon$
By $(1)$, there exist:
- $A_1, \ldots, A_n \in \AA_0$
such that:
- $\ds (3): \quad \forall k \in \set {1, \ldots, n} : \map \mu {A_k \symdif B_k} \le \paren {\frac \epsilon {\size {n \alpha _k} } }^p$
By definition of $\AA_0$, for each $k$ there exists an $n_k \ge 0$ such that:
- $A_k \in \FF_{n_k}$
Let $N := \max \set {n_1, \ldots, n_k}$.
Then:
- $A_1, \ldots, A_n \in \FF_N$
In particular:
- $\ds \sum_{k \mathop = 1}^n \alpha_k \chi_{A_k} \in \map {L^p} {X, \FF_N, \mu}$
Moreover:
\(\ds \norm {f - \sum_{k \mathop = 1}^n \alpha_k \chi_{A_k} }_p\) | \(=\) | \(\ds \norm {f - \sum_{k \mathop = 1}^n \alpha_k \chi_{B_k} + \sum_{k \mathop = 1}^n \alpha_k \paren {\chi_{B_k} - \chi_{A_k} } }_p\) | ||||||||||||
\(\ds \) | \(\le\) | \(\ds \norm {f - \sum_{k \mathop = 1}^n \alpha_k \chi_{B_k} }_p + \sum_{k \mathop = 1}^n \norm {\alpha_k \paren {\chi_{B_k} - \chi_{A_k} } }_p\) | ||||||||||||
\(\ds \) | \(=\) | \(\ds \norm {f - \sum_{k \mathop = 1}^n \alpha_k \chi_{B_k} }_p + \sum_{k \mathop = 1}^n \size {\alpha_k} \norm { \chi_{B_k} - \chi_{A_k} }_p\) | ||||||||||||
\(\ds \) | \(=\) | \(\ds \norm {f - \sum_{k \mathop = 1}^n \alpha_k \chi_{B_k} }_p + \sum_{k \mathop = 1}^n \size {\alpha_k} \paren {\int \size {\chi_{B_k} - \chi_{A_k} }^p \rd \mu }^{\frac 1 p}\) | Definition of $L^p$-norm | |||||||||||
\(\ds \) | \(=\) | \(\ds \norm {f - \sum_{k \mathop = 1}^n \alpha_k \chi_{B_k} }_p + \sum_{k \mathop = 1}^n \size {\alpha_k} \paren {\int \chi_{A_k \symdif B_k} \rd \mu }^{\frac 1 p}\) | ||||||||||||
\(\ds \) | \(=\) | \(\ds \norm {f - \sum_{k \mathop = 1}^n \alpha_k \chi_{B_k} }_p + \sum_{k \mathop = 1}^n \size {\alpha_k} \map \mu {A_k \symdif B_k}^{\frac 1 p}\) | Definition of Integral of Positive Simple Function | |||||||||||
\(\ds \) | \(\le\) | \(\ds \epsilon + \sum_{k \mathop = 1}^n \size {\alpha_k} \frac \epsilon {\size {n \alpha _k} }\) | by $(2)$ and $(3)$ | |||||||||||
\(\ds \) | \(=\) | \(\ds 2 \epsilon\) |
This needs considerable tedious hard slog to complete it. In particular: Details To discuss this page in more detail, feel free to use the talk page. When this work has been completed, you may remove this instance of {{Finish}} from the code.If you would welcome a second opinion as to whether your work is correct, add a call to {{Proofread}} the page. |