Take the 2-minute tour ×
Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

I'm trying to check wheter this function series is uniformly converges or not:$f_n(x)=\frac{2nx}{1+n^2x^2}$.

I tried to check $\lim_{n \to \infty} \sup |f_n(x)-f(x)|$ is 0 where $f(x)=\lim_{n \to \infty}f_n(x)$.

Let $r_n(x)=f_n(x)-f(x)$. I checked maximum points by $r_n'(x)=0$ and got $x=\frac{1}{n}$ as maximim, so $r_n(\frac{1}{n})=1$ so $\lim_{n \to \infty} \sup |f_n(x)-f(x)|=1 \neq0$ so I concluded that there's not uniform convergence. But then by simple inequalities: $|f_n(x)-f(x)|=\frac{2nx}{1+n^2x^2}\leq \frac{2}{\frac {1}{nx}+nx}\leq\frac{2}{n}\to0$, so there is a uniform convergence. whats wrong with my first try?

Edit:$x \in [1, \infty)$ Thanks a lot.

share|improve this question

1 Answer 1

up vote 3 down vote accepted

If you study the convergence on the interval $[1,\infty)$, then you're interested in $\sup\limits_{x\in[1,\infty)} |f_n(x)-f(x)|$. Since $\frac1n\notin[1,\infty)$, this is not what you computed in the first part.

I guess that if you have a look at your computations again, you'll find out that $r'_n(x)<0$ for $x>\frac1n$, which implies that the function $r_n(x)=f_n(x)-f(x)$ is decreasing on $[1,\infty)$. (In fact, it is decreasing already on a bigger interval $[1/n,\infty)$.)

Thus the maximal possible value is at the leftmost point of this interval, which is $r_n(1)=f_n(1)=\frac{2n}{1+n^2}$. This converges to zero.

So your first approach is fine, the only thing is you have to find the supremum of $|f_n(x)-f(x)|$ on the same interval, for which you want to test the uniform convergence.


EDIT: The rest is about the first version of the post, without the condition $x\in[1,\infty)$. I think the best thing is to keep it here too.

In your second solution you used the inequality which is not valid. You argued that: $$\frac1{nx}+nx \ge n \Rightarrow \frac2{\frac1{nx}+nx} \le \frac 2n.$$

In fact you only have $$\frac1{nx}+nx \ge 2 \Rightarrow \frac2{\frac1{nx}+nx} \le \frac 22=1,$$ which is consistent with your first solution.


There are many ways to see that for $t>0$ we have $$t+\frac1t\ge 2.$$ (This is the inequality I used above for $t=nx$.)

E.g. you can study the derivative of the function $f(t)=t+\frac1t$. Quite simple way to see this is using AM-GM inequality, which gives you: $$\frac{t+\frac1t}2\ge \sqrt{t\cdot\frac1t}=1.$$

It is known that we get the equality in AM-GM only when the summands are equal, which is in our case $t=\frac1t$, i.e. $nx=t=1$ and $x=\frac1n$; just the same as you found in your first approach.

share|improve this answer
    
I forgot to mention that $x \in [1, \infty)$, I'm sorry. –  Jozef Jan 28 '12 at 11:39

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

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