3

Let $f(x)$ be continuous for all $\mathbb R$. $$\lim_{x\to\infty}f(x)=L_1$$ and $$\lim_{x\to-\infty}f(x)=L_2$$ Where $L_1,L_2$ belong to $\mathbb R$.

Prove that $f(x)$ is bounded for all $\mathbb R$.

My problem with this conjecture: Isn't $f(x)$=$1/x$ a counterexample? In this case, $L_1,L_2=0$ and the function is not bounded. Did I miss something?

Thanks in advance.

tomasz
  • 35,474
pie
  • 145

5 Answers5

5

Hint: It is likely that you already have a theorem to the effect a function which is continuous on a closed (bounded) interval is bounded on that interval. Perhaps it is in the form that such a function attains a maximum and a minimum.

Let $A$ be a negative number such that $f(x)$ is not far from $L_2$ for $x\lt A$, and let $B$ be a positive number such that $f(x)$ is not far from $L_1$ for $x\gt B$.

André Nicolas
  • 507,029
3

Proof: $$\exists m<0: x<m\implies \left|f(x)-L_1\right|<1\iff L_1-1<f(x)<L_1+1$$ and $$\exists M>0: x>M\implies \left|f(x)-L_2\right|<1\iff L_2-1<f(x)<L_2+1$$ Now in $[m,M]$, by continuity $f$ is bounded. Can you finish this off?

Nameless
  • 13,456
  • So if we take K=$max[$L_1+1,L_2 + 2$]$ to be the upper boundry and k=$min[$L_1+1,L_2 + 2$]$ to be the lower boundry, it is obvious, that for P=max[K,k], |f(x)|<P ? – pie Dec 29 '12 at 15:27
  • The limits of f(x) when x->-∞ and x->∞. – pie Dec 29 '12 at 15:29
  • My bad, meant L2+1.

    P.S. How do I math-type in the comments?

     – pie Dec 29 '12 at 15:31
  • Sorry, I didn't manage to activate the chat, to my emberassment. – pie Dec 29 '12 at 15:47
  • @pie: mathmode works in comments pretty much the same as everywhere else, just use $ signs. – tomasz Dec 29 '12 at 15:51
  • So why is saying that $f(x)$ is bounded by $max[L1+1,L2+1]$ from above and by $min[L1-1,L2-1]$ from below incorrect? And what's the importance of $f$ being bounded in [m,M]? – pie Dec 29 '12 at 15:53
  • @pie You had written $L_1+1$ and not $L_1-1$ for the lower bound. Now $f$ is bounded by these numbers in $(-\infty,m)\cup (M,\infty)$ But we have to prove that itis bounded in $[m,M]$ don't we? – Nameless Dec 29 '12 at 15:55
2

A different proof: by the existence of the limits, the function $f$ can be extended to a continuous function on $[-\infty,+\infty]$, which is compact (you can compose $f$ with $\tan$ and think instead of it as a function on $(-\pi/2,+\pi/2)$ extended to $[-\pi/2,\pi/2]$ if you're not comfortable with $[-\infty,+\infty]$).

A continuous image of a compact set is compact (or: a continuous function on a closed interval is bounded and attains its extremes, in the language of elementary analysis), so the extension of $f$ is bounded, and so is $f$.

tomasz
  • 35,474
1

by existence of 2 limits,f must be bounded in some neighbourhood of infinity and minus infinity say[-infinity,a]and [b,infinity] by some M.on [a,b] again f must be bounded by continity by N say.hence f must be bounded by max(M,N)

Koushik
  • 4,472
1

Given the two limits, we can find $x_1<x_2$ such that $m:=\min(L_1,L_2)-1\leq f(x) \leq \max(L_1,L_2)+1=:M$ for all $x\leq x_1$ and all $x\geq x_2$.

Now $f$ is continuous on the compact interval $[x_1,x_2]$, hence there exist $m'\leq M'$ such that $m'\leq f(x)\leq M'$ for all $x$ in $[x_1,x_2]$.

Finally, we see that $f$ is bounded above by $\max(M,M')$ and below by $\min(m,m')$ on $\mathbb{R}$.

Julien
  • 44,791