Given a Cauchy sequence of functions $\{f_n(x)\}$, it is obvious that it converges pointwise to some function $f(x)$. To prove uniform convergence, the standard argument goes as follows:

Let $\varepsilon>0$. Choose $N$ such that for all $m,n \geq N$, $|f_n(x)-f_m(x)|<\varepsilon$. Fix $n$ and take $m \to \infty$. Since $f_m(x) \to f(x)$, it follows that $|f_n(x)-f(x)| \leq \varepsilon$ for all $n \geq N$ and $x \in E$.

I don't find the "take the limit of $m$ to infinity" part rigorous enough because $f_m(x) \to f(x)$ is only pointwise and depends on $x$. How can we derive a conclusion of for all $x \in E$ then? What am I missing here? It also sounds a bit like heuristics and not really Analysis. How to make the argument formal?

  • $\begingroup$ What is the norm for your functional space? Sup norm? $\endgroup$ – Graham L Mar 30 at 9:18

First you pick $N$ such that for all $n,m>N$ and for all $x$ we have (here we're using that $(f_n)_{n\in \mathbb{N}})$ is a Cauchy sequence with respect to the supremum norm) $$ \vert f_n(x)-f_m(x) \vert < \varepsilon/2.$$ Now fix some $x$ and pick $m(x)>n>N$ such that $$ \vert f_{m(x)} (x) - f(x)\vert < \varepsilon/2$$ this you can do as $f_n(x) \rightarrow f(x)$. By the triangle inequality we get $$ \vert f_n(x)- f(x)\vert \leq \vert f_n(x)- f_{m(x)} (x) \vert + \vert f_{m(x)}(x) - f(x)\vert < \varepsilon/2 + \varepsilon/2 $$ Hence we get for every $\varepsilon >0$ some $N$ such that for all $n>N$ $$ \sup_{x\in E}\ \vert f_n(x) -f(x) \vert < \varepsilon.$$ Thus, $(f_n)_{n\in \mathbb{N}}$ converges uniformly to $f$.


Your Answer

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

Not the answer you're looking for? Browse other questions tagged or ask your own question.