Let $(X,d_X)$ and $(Y,d_Y)$ be two metric spaces, and let $\{f_n\}$ be a sequence of functions $f_n: X \rightarrow Y$. For any function $f : X \rightarrow Y$ the following are equivalent:

(i) $\{f_n\}$ converges uniformly to $f$

(ii) $\sup\{d_Y(f_n(x),f(x)) \ | \ x \in X\} \rightarrow 0$ as $n$ $\rightarrow \infty$

Proof: (i) $\Rightarrow$ (ii): Assume $\{f_n\}$ converges uniformly to $f$. $\forall \epsilon > 0$ $\exists N \in \mathbb{N}$ s.t. $d_Y(f_n(x),f(x)) < \epsilon$ $\forall x \in X$ and $\forall n \geq N$. This means that $\sup\{d_Y(f_n(x),f(x)) \ | \ x \in X\} \leq \epsilon$ $\forall n \geq N$, and since $\epsilon$ is arbitrary, this implies $\sup\{d_Y(f_n(x),f(x)) | x \in X\} \rightarrow 0$

I don't understand why there is $\leq$ and not $<$ in $\sup\{d_Y(f_n(x),f(x)) \ | \ x \in X\} \leq \epsilon$. Does it follow from the fact that a supremum is the smallest element greater than or equal to the others?


If $A\subseteq \mathbb{R}$ is such that any $a\in A$ satisfies $a<b$ then $\sup A \leq b$.

The weak inequality is necessary. For instance if $A=(0,1)$, then any $a\in A$ satisfies $a<1$ however $\sup A = 1$.

This should explain the weak inequality in your question. However since $\varepsilon$ is arbitrary it doesn't really change anything.


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.