next up previous contents index
Next: (Rates of growth) Up: Differentiability Previous: l'Hôpital revisited   Contents   Index

Infinite limits

We can use Proposition 4.28 to get results about infinite limits.

Example 5.26   Evaluate $\displaystyle \lim_{{x \to \infty}}^{}$x log$\displaystyle \left(\vphantom{1 + \frac{1}{ x}}\right.$1 + $\displaystyle {\frac{{1}}{{x}}}$$\displaystyle \left.\vphantom{1 + \frac{1}{ x}}\right)$.

Solution. 

$\displaystyle \lim_{{x \to \infty}}^{}$x log$\displaystyle \left(\vphantom{1 + \frac{1}{ x}}\right.$1 + $\displaystyle {\frac{{1}}{{x}}}$$\displaystyle \left.\vphantom{1 + \frac{1}{ x}}\right)$ = $\displaystyle \lim_{{x \to \infty}}^{}$$\displaystyle {\frac{{\log(1+y)}}{{ y}}}$    writing y = 1/x,      
  = $\displaystyle \lim_{{y \to 0+}}^{}$$\displaystyle {\frac{{\log(1+y)}}{{ y}}}$  
  = $\displaystyle \lim_{{y \to 0}}^{}$$\displaystyle {\frac{{\log(1+y)}}{{ y}}}$ = 1.  

The last step is valid, since the final limit exists by l'Hôpital; note also that this gives another way of finding an = (1 + 1/n)n.


Exercise 5.27   Evaluate $\displaystyle \lim_{{x \to \infty}}^{}$x sin$\displaystyle \left(\vphantom{\frac{1}{ x}}\right.$$\displaystyle {\frac{{1}}{{x}}}$$\displaystyle \left.\vphantom{\frac{1}{ x}}\right)$.

Proposition 5.28 (l'Hôpital's rule: infinite limits)   Let f and g be functions such that $ \lim_{{x\to \infty}}^{}$f (x) = $ \lim_{{x\to \infty}}^{}$g(x) = $ \infty$, and suppose that $\displaystyle \lim_{{x \to \infty}}^{}$$\displaystyle {\frac{{f'(x)}}{{g'(x)}}}$ exists. Then

$\displaystyle \lim_{{x \to \infty}}^{}$$\displaystyle {\frac{{f(x)}}{{g(x)}}}$ = $\displaystyle \lim_{{x \to \infty}}^{}$$\displaystyle {\frac{{f'(x)}}{{g'(x)}}}$.

Proof. (Sketch for interest -- not part of the course). Pick $ \epsilon$ > 0 and choose a such that

$\displaystyle \lim_{{x \to \infty}}^{}$$\displaystyle \left\vert\vphantom{\frac{f'(x)}{g'(x)} - l}\right.$$\displaystyle {\frac{{f'(x)}}{{g'(x)}}}$ - l$\displaystyle \left.\vphantom{\frac{f'(x)}{g'(x)} - l}\right\vert$ < $\displaystyle \epsilon$    for all x > a.    

Then pick K such that if x > K, then g(x) - g(a)$ \ne$ 0. By Cauchy,

$\displaystyle {\frac{{f'(c)}}{{g'(c)}}}$ = $\displaystyle {\frac{{f(x) - f(a)}}{{g(x) - g(a)}}}$    for all x > K.    

Note that although c depends on x, we always have c > a. Then
$\displaystyle {\frac{{f(x)}}{{g(x)}}}$ = $\displaystyle {\frac{{f(x) - f(a)}}{{g(x) - g(a)}}}$.$\displaystyle {\frac{{f(x)}}{{f(x) - f(a)}}}$.$\displaystyle {\frac{{g(x) - g(a)}}{{g(x)}}}$,  
  $\displaystyle \to$ l.1.1    as x$ \to$$ \infty$.      

$ \qedsymbol$



Subsections
next up previous contents index
Next: (Rates of growth) Up: Differentiability Previous: l'Hôpital revisited   Contents   Index
Ian Craw 2002-01-07