Let $f_n:[-1, 1] \to \mathbb{R}$ defined by $f_n(x) = \dfrac{x}{1 + n^2 x^2}$ be a sequence of functions where $n \in \mathbf{N}$.
Goal: Show that $f_n$ converges uniformly to a function by showing that $|f_n(x)| \leq \dfrac{1}{n}$.
I am able to bound it up to some point. I just couldn't show that it is precisely bounded above by $1/n$. My attempt so far:
Proof.
$f_n$ clearly converges pointwise to $0$. Now, we check uniform convergence.
If $x = 0$, then $f_n \to 0$ uniformly (because it is identically $0$) for $x \in [-1, 1]$.
Now suppose $x \neq 0$. Then,
$$|f_n(x) - 0| = \left|\dfrac{x}{1 + n^2 x^2}\right| \leq \left| \dfrac{1}{n^2 x^2}\right| \leq \left| \dfrac{1}{n^2 x}\right| = \dfrac{1}{n^2 |x|}$$
End of attempt.
Here is where I got stuck. I can see that if we can bound $f_n(x)$ above by $1/n$, then the following has to be true:
$$ \dfrac{1}{n^2 |x|} \leq \dfrac{1}{n}\iff |x| \geq \dfrac{1}{n} \iff n|x| \geq 1$$
But, I just can't see how the above is true. Am I missing something?
Observe that, when $\lvert x \rvert >0$, $$\lvert f_n(x)\rvert =\frac{1}{2n}M_{-1}\left(n\lvert x \rvert,\tfrac{1}{n\lvert x \rvert}\right)$$ where $M_{-1}$ denotes the harmonic mean. Let $M_0$ denote the geometric mean. the geometric mean–harmonic mean inequality states that $$M_{-1}\leq M_0\text{,}$$ with equality holding iff all operands are equal. Consequently, $$M_{-1}\left(n\lvert x \rvert,\tfrac{1}{n\lvert x \rvert}\right)\leq 1\text{,}$$ with equality holding iff $\lvert x \rvert =\tfrac{1}{n}$. Hence $$\lvert f_n(x)\rvert\leq \tfrac{1}{2n}\text{.}$$