Confirm if this is correct: Proof of convergence with respect to the Schwartz metric implies convergence in Schwartz space

111 Views Asked by At

This is the definition of convergence in Schwartz space:

We say that $f_n \rightarrow f$ in $S$ if $\lim_{n \to \infty} ||f-f_n||_{ \alpha , \beta} \rightarrow 0$ for all $ \alpha , \beta \in \mathbb{N_{0}^{d}}$

Where,

The Schwartz space $S( \mathbb{R^d}) $ is the set of functions $f \in C^{\infty}(\mathbb{R^d})$ such that:

$$||f||_{ \alpha , \beta} = || x^{\alpha} \partial_{x}^{\beta} f||_{\infty} < \infty$$

for all multiindices $ \alpha , \beta$.

Proposition: ($S$ is a metric space). Convergence in $S$ is equivalent to convergence with respect to the metric:

$$d_{S}(f,g) = \sum_{n=0}^{\infty} 2^{-n} \sup_{ | \alpha| + | \beta | = n} \frac{ ||f-g||_{ \alpha , \beta}}{1+||f-g||_{\alpha,\beta}}$$

Where,

The Schwartz metric,

Now, maybe I'm confused or probably overthinking, but when I'm thinking 'convergence' in $d_{S}$, this is independent on $n$ (the index of the sum) Please correct me if I'm wrong.

However, suppose $d_{S}(f_{j},f) \rightarrow 0$, since each term of the sum is positive, all of the terms have to approach zero, which means that $||f_j -f||_{\alpha , \beta} \rightarrow 0$ $\forall \alpha , \beta$ since the $sup$ itself approaches $0$

Now the notes do not use this method of proof which is supposed to be trivial, so I'm starting to suspect that there's something wrong with my reasoning.

1

There are 1 best solutions below

4
On

Your problem has not much to do with the Schwartz space. For an increasing sequence $p_n$ of seminorms on a vector space $X$ you want to show that a sequence $(x_k)_{k\in\mathbb N}$ converges with respect to each $p_n$ to a vector $x$ if and only if it converges with respect to $d(x,y)=\sum\limits_{n=1}^\infty 2^{-n} \frac{p_n(x-y)}{1+p_n(x-y)}$. The crucial point is that the condition $d(x,y)<\varepsilon$ only depends on the first $n(\epsilon)$ terms of the series. It seems that this form of the metric was first used by Fréchet who also introduced another one, namely $$D(f,g)=\sup\{p_n(x-y) \wedge 1/n: n\in\mathbb N\} $$ where $a\wedge b$ denotes the minimum of $a,b\ge 0$. Here it is much easier to see that $D(x,y)<1/n$ if and only if $p_n(x-y)< 1/n$ (to check topological concepts, it is enough to consider $\varepsilon=1/n$).