
Calculus Calculus Math Forum 
 LinkBack  Thread Tools  Display Modes 
February 18th, 2015, 10:20 AM  #1 
Newbie Joined: Feb 2015 From: United Kingdom Posts: 16 Thanks: 1  Propositions concerning a differentiable function.
I am trying to solve the following problems: Suppose that f: (0, ∞) → R is differentiable. Determine whether each of the following statements is true or false: (a) If limx→∞ f(x) exists and is finite, and limx→∞ f'(x) = b, then b = 0. (b) If limx→∞ f(x) exists and is finite, then limx→∞ f'(x) = 0. (c) If limx→∞ f'(x) = 0, then limx→∞ f(x) exists. (d) If limx→∞ f'(x) = 0, then limx→∞ f(x)/x = 0. I have drawn pictures for (a), and I think that it is true. However, I'm not sure how to prove this. I've managed to find a counterexample for (c). I'm not sure how to approach (b) and (d), although I have noticed that (a) and (b) are related. Could I have suggestions on how to approach these questions? 
February 18th, 2015, 11:27 AM  #2 
Math Team Joined: Dec 2013 From: Colombia Posts: 7,559 Thanks: 2561 Math Focus: Mainly analysis and algebra 
a) Suppose that $b \ne 0$. Consider the $\delta\epsilon$ definition of the limit of $f$ and come to a contradiction. b) What if the limit doesn't exist? 
February 18th, 2015, 05:54 PM  #3 
Math Team Joined: Dec 2013 From: Colombia Posts: 7,559 Thanks: 2561 Math Focus: Mainly analysis and algebra 
Recall that $$\lim_{x \to \infty} f'(x) = b$$means that for all $\epsilon \gt 0$ there exists a $\delta$ such that $$x \gt \delta \implies f'(x)  b \lt \epsilon$$ Thus we have $$\newcommand{\dt}{\,\mathrm d t}\begin{aligned} && \epsilon &\lt f'(t)  b \lt \epsilon &\text{for all $t \gt \delta$} \\ && \int_\delta^x \epsilon \dt &\lt \int_\delta^x f'(t)  b \dt \lt \int_\delta^x \epsilon \dt \\ && (\delta  x)\epsilon &\lt f(x)  f(\delta)  b(x  \delta) \lt (x  \delta)\epsilon \\ &\text{gathering all the constant terms into one, we get} & (b\epsilon) x &\lt f(x)  L \lt (b + \epsilon)x &\text{(equation 1)}\end{aligned}$$ Now we address part a) When $b \gt 0$ we take the left hand inequality of equation 1, with $\epsilon = {b \over 2}$ (this sets the value of $L$, but it is still a constant) to get $${b \over 2}x + L \lt f(x)$$and since the left hand side grows without bound, so does $f(x)$. This means that $f(x)$ has no limit, contradicting the assumptions of proposition a). When $b \lt 0$ we take the right hand inequality of equation 1, with $\epsilon = {b \over 2}$ to get $$f(x) \lt L + {b \over 2}x$$and since $b \lt 0$ the right hand side decreases without bound and therefore $f(x)$ does too contradicting the assumptions of proposition a). Thus $$\lim_{x \to \infty} f'(x) = b \implies b = 0$$ Proposition b) omits the assumption that the limit exists. So we may search for a counter example where there is no limit. $$f(x) = c + {1 \over x} \sin x^2 \implies \lim_{x \to \infty} f(x) = c \\ f'(x) = {1 \over x}2x \cos x^2 {1 \over x^2}\sin x^2 = 2\cos x^2 {1 \over x^2}\sin x^2$$ The second term of the derivative converges to zero, but the first oscillates between 0 and 1. So $f'(x)$ does not tend to a limit as $x \to \infty$. For proposition c) $f'(x) = {1 \over x}$ is a counterexample. For proposition d) we return to equation 1 and divide by $x$ to get $$b  \epsilon \lt {f(x) \over x}  {L \over x} \lt b + \epsilon \\ {L \over x} \epsilon \lt {f(x) \over x}  b \lt {L \over x} + \epsilon$$ And this is valid for all $x \gt \delta$ so ${L \over x} \lt {L \over \delta}$ and thus we have $$\left{f(x) \over x}  b\right \lt \left{L \over \delta}\right + \epsilon$$ both terms on the right can be made arbitrarily small by choosing small enough $\epsilon$ and large enough $\delta$ so we have the convergence of ${f(x) \over x}$ as required (note that $b=0$ in the particular case given). 
February 19th, 2015, 10:45 PM  #4 
Math Team Joined: Dec 2013 From: Colombia Posts: 7,559 Thanks: 2561 Math Focus: Mainly analysis and algebra 
There are some errors in the work above (although the ideas are correct). In particular equation (1) is incorrect. Recall that $$\lim_{x \to \infty} f'(x) = b$$means that for all $\epsilon \gt 0$ there exists a $\delta$ such that $$x \gt \delta \implies f'(x)  b \lt \epsilon$$ Thus we have $$\newcommand{\dt}{\,\mathrm d t}\begin{aligned} && \epsilon &\lt f'(t)  b \lt \epsilon &\text{for all $t \gt \delta$} \\ && \int_\delta^x \epsilon \dt &\lt \int_\delta^x f'(t)  b \dt \lt \int_\delta^x \epsilon \dt \\ && (\delta  x)\epsilon &\lt f(x)  f(\delta)  b(x  \delta) \lt (x  \delta)\epsilon &\text{(equation 1)}\end{aligned}$$ When $b \gt 0$ we take the left hand inequality of equation (1), with $\epsilon = {b \over 2}$ to get $${b \over 2}x + L \lt f(x) \qquad (L = f(\delta)  \tfrac{b}{2}\delta \quad \text{a constant})$$and since the left hand side grows without bound, so does $f(x)$. This means that $f(x)$ has no limit, contradicting the assumptions of proposition a). And similarly the right hand inequality shows that $b \lt 0$ implies that $f(x)$ is unbounded below. For proposition d) we return to equation (1), taking $b=0$, and divide by $x$ to get $$\left({\delta \over x}  1\right)\epsilon \lt {f(x)  f(\delta) \over x} \lt \left(1  {\delta \over x}\right)\epsilon$$ Since the equation is valid for $x \gt \delta$ we have ${\delta \over x} \lt 1$ and so we can write $1  {\delta \over x} > p > 0$ for large enough $x$ and we can thus write $p\epsilon = \epsilon'$ to get $$\begin{aligned} \epsilon' \lt {f(x)  f(\delta) \over x} &\lt \epsilon' \\ \implies \left {f(x)  f(\delta) \over x} \right &\lt \epsilon' \\ \text{so} \quad \lim_{x\to \infty} {f(x)  f(\delta) \over x} &= 0 \\ \lim_{x\to \infty} \left({f(x) \over x}  {f(\delta) \over x}\right) &= 0 \\ \lim_{x\to \infty} {f(x) \over x}  \lim_{x\to \infty} {f(\delta) \over x} &= 0 \\ \lim_{x\to \infty} {f(x) \over x} &= \lim_{x\to \infty} {f(\delta) \over x} = 0 \\ \end{aligned}$$ as required. 

Tags 
calculus analysis limits, differentiable, function, propositions 
Search tags for this page 
Click on a term to search for related topics.

Thread Tools  
Display Modes  

Similar Threads  
Thread  Thread Starter  Forum  Replies  Last Post 
Differentiable function  bereprada  Calculus  4  March 10th, 2014 08:38 PM 
differentiable function????  ABHISHEK MEENA  Calculus  1  December 26th, 2012 11:12 AM 
Probability and Expectation of Logical Propositions  restin84  Advanced Statistics  0  February 28th, 2012 08:56 PM 
differentiable function?  Camille91  Real Analysis  2  November 17th, 2011 09:22 AM 
no differentiable function.....  rose3  Real Analysis  1  January 14th, 2010 05:50 AM 