Show that f'(0) exists and find its value

  • Context: MHB 
  • Thread starter Thread starter evinda
  • Start date Start date
  • Tags Tags
    Value
Click For Summary
SUMMARY

The discussion centers on proving the existence of the derivative f'(0) for a function f: (-a,a) → ℝ that is continuous at 0, under the condition that lim_{x→0} (f(x) - f(tx))/x = b for t in (0,1). Participants concluded that f'(0) exists and is equal to b/(1-t). The use of the Mean Value Theorem and L'Hôpital's Rule was debated, with emphasis on the necessity of proving differentiability at 0 before applying these theorems.

PREREQUISITES
  • Understanding of limits and continuity in calculus.
  • Familiarity with the Mean Value Theorem.
  • Knowledge of L'Hôpital's Rule.
  • Basic concepts of differentiability and its implications.
NEXT STEPS
  • Study the Mean Value Theorem and its applications in proving differentiability.
  • Learn about L'Hôpital's Rule and its conditions for use.
  • Explore the implications of continuity on differentiability in real analysis.
  • Investigate examples of functions that are continuous but not differentiable at specific points.
USEFUL FOR

Mathematics students, calculus instructors, and anyone interested in the analysis of functions and their derivatives, particularly in the context of continuity and differentiability.

evinda
Gold Member
MHB
Messages
3,741
Reaction score
0
Hello! :D
I am looking at this exercise:
Let f: (-a,a)\rightarrow \mathbb{R} continuous at 0.We suppose that for a t \epsilon (0,1) this condition is satisfied:

\lim_{x\rightarrow 0}\frac{f(x)-f(tx)}{x}=b

Show that f'(0) exists and find its value.

I thought that I could write it like that \lim_{x\rightarrow 0}\frac{f(x)-f(0)+f(0)-f(tx)}{x}=b \Rightarrow \lim_{x\rightarrow 0}\frac{f(x)-f(0)}{x}+\lim_{x\rightarrow 0}\frac{f(0)-f(tx)}{x}=b\Rightarrow (1-\frac{1}{t})\lim_{x\rightarrow 0}\frac{f(x)-f(0)}{x}=b
So,f'(0) exists and is equal to \frac{bt}{t-1} .Are my thoughts right or am I wrong?
 
Physics news on Phys.org
If the first derivative exists at $0$ then by L'Hosiptal rule we have

$$ \lim_{x\rightarrow 0}\frac{f(x)-f(tx)}{x}= f'(0)-tf'(0) = b $$

$$f'(0)= \frac{b}{1-t}$$
 
ZaidAlyafey said:
If the first derivative exists at $0$ then by L'Hosiptal rule we have

$$ \lim_{x\rightarrow 0}\frac{f(x)-f(tx)}{x}= f'(0)-tf'(0) = b $$

$$f'(0)= \frac{b}{1-t}$$

And that is the only thing I have to say?So...because of the fact that $$ \frac{b}{1-t}\epsilon \mathbb{R}$$ ,$$ f'(0) $$ exists,and its value is $$ \frac{b}{1-t} $$ ?
 
evinda said:
Hello! :D
I am looking at this exercise:
Let f: (-a,a)\rightarrow \mathbb{R} continuous at 0.We suppose that for a t \epsilon (0,1) this condition is satisfied:

\lim_{x\rightarrow 0}\frac{f(x)-f(tx)}{x}=b

Show that f'(0) exists and find its value.

I thought that I could write it like that \lim_{x\rightarrow 0}\frac{f(x)-f(0)+f(0)-f(tx)}{x}=b \Rightarrow \lim_{x\rightarrow 0}\frac{f(x)-f(0)}{x}+\lim_{x\rightarrow 0}\frac{f(0)-f(tx)}{x}=b\Rightarrow (1-\frac{1}{t})\lim_{x\rightarrow 0}\frac{f(x)-f(0)}{x}=b
So,f'(0) exists and is equal to \frac{bt}{t-1} .Are my thoughts right or am I wrong?

The mean value theorem extablishes that, given a function f(*) continuous in [a,b] and differentiable in (a,b), it exists a number c in (a,b) such that... $\displaystyle f^{\ '} (c) = \frac{f(b) - f(a)} {b - a}\ (1)$

Setting in (1) b=x and a= t x, if ...

$\displaystyle \lim_{ x \rightarrow 0} \frac{f(x) - f(t\ x)}{x} = \alpha\ (2)$

... is also...

$\displaystyle \lim_{ x \rightarrow 0} \frac{f(x) - f(t\ x)}{x\ (1 - t)} = \frac{\alpha}{1 - t} = f^{\ '}(0)\ (3)$

View attachment 1784

Merry Christmas from Serbia

$\chi$ $\sigma$
 

Attachments

  • eb191d59c14248f1e362dd5eb1d3102c.jpg
    eb191d59c14248f1e362dd5eb1d3102c.jpg
    42.1 KB · Views: 108
chisigma said:
The mean value theorem extablishes that, given a function f(*) continuous in [a,b] and differentiable in (a,b), it exists a number c in (a,b) such that... $\displaystyle f^{\ '} (c) = \frac{f(b) - f(a)} {b - a}\ (1)$

Setting in (1) b=x and a= t x, if ...

$\displaystyle \lim_{ x \rightarrow 0} \frac{f(x) - f(t\ x)}{x} = \alpha\ (2)$

... is also...

$\displaystyle \lim_{ x \rightarrow 0} \frac{f(x) - f(t\ x)}{x\ (1 - t)} = \frac{\alpha}{1 - t} = f^{\ '}(0)\ (3)$

And which interval [a,b] do we take??
 
evinda said:
And which interval [a,b] do we take??

If You chose a= t x and b = x, then b-a = x (1 - t) and t x < c < x. If x tends to 0, then a, b and c tend to 0 and starting from...

$\displaystyle \lim_{x \rightarrow 0} \frac{f(x) - f(t\ x)}{x} = \alpha\ (1)$

... You arrive at...

$\displaystyle f^{\ '} (0) = \frac{\alpha}{1 - t}\ (2)$

View attachment 1785

Merry Christmas from Serbia

$\chi$ $\sigma$
 

Attachments

  • eb191d59c14248f1e362dd5eb1d3102c.jpg
    eb191d59c14248f1e362dd5eb1d3102c.jpg
    42.1 KB · Views: 116
chisigma said:
If You chose a= t x and b = x, then b-a = x (1 - t) and t x < c < x. If x tends to 0, then a, b and c tend to 0 and starting from...

$\displaystyle \lim_{x \rightarrow 0} \frac{f(x) - f(t\ x)}{x} = \alpha\ (1)$

... You arrive at...

$\displaystyle f^{\ '} (0) = \frac{\alpha}{1 - t}\ (2)$

And [a,b]=[tx,x] should be \subseteq (-a,a) .Right?
 
evinda said:
And [a,b]=[tx,x] should be \subseteq (-a,a) .Right?

So,can I start like that:f is continuous in [a,b]\subseteq (-a,a) and differentiable in (a,b),so it exists a number c in (a,b) such that f&#039;(c)=\frac{f(b)-f(a)}{b-a}.
We select a=tx and b=x . ?
 
evinda said:
Let f: (-a,a)\rightarrow \mathbb{R} continuous at 0.We suppose that for a t \epsilon (0,1) this condition is satisfied:

\lim_{x\rightarrow 0}\frac{f(x)-f(tx)}{x}=b

Show that f&#039;(0) exists and find its value.
ZaidAlyafey's comment shows that if $f'(0)$ exists then it is equal to $\dfrac b{1-t}.$ The snag with this, and with all the other comments, is that you are not told that $f$ is differentiable. You are not even told that it is continuous, except at the single point $0$. So you cannot use l'Hôpital's rule, or the mean value theorem, or any other result that requires differentiability, and you need to prove that $f$ is differentiable at $0$.

You are told that \lim_{x\rightarrow 0}\frac{f(x)-f(tx)}{x}=b. So given $\varepsilon>0$ there exists $\delta>0$ such that $$\left|\frac{f(x)-f(tx)}{x} - b\right| < \varepsilon$$ whenever $|x|<\delta.$ Therefore $$|f(x) - f(tx) - bx| < \varepsilon|x|.$$ Next, if $|x|<\delta$ then $|t^nx|<\delta$ for all $n\geqslant1$ (because $t<1$). So in the above inequality we can replace $x$ by $tx$, $t^2x,\ldots, t^{n-1}x,$ to get $$|f(tx) - f(t^2x) - btx| < \varepsilon t|x|,$$ $$|f(t^2x) - f(t^3x) - bt^2x| < \varepsilon t^2|x|,$$ $$\ldots$$ $$|f(t^{n-1}x) - f(t^nx) - bt^{n-1}x| < \varepsilon t^{n-1}|x|.$$ Now add those inequalities and use the triangle inequality, to get $$\left|\sum_{k=0}^{n-1}\bigl(f(t^kx) - f(t^{k+1}x) - bt^kx\bigr)\right| \leqslant \sum_{k=0}^{n-1}\bigl|f(t^kx) - f(t^{k+1}x) - bt^kx\bigr| < \sum_{k=0}^{n-1} \varepsilon t^k|x|.$$ It follows, by summing the geometric and telescoping series, that $$\left| f(x) - f(t^nx) - \frac{b(1-t^n)x}{1-t} \right| < \frac{\varepsilon(1-t^n)|x|}{1-t}.$$ Now you can let $n\to\infty$ and use the facts that $t^nx\to0$ and $f$ is continuous at $0$, to deduce that $$\left| f(x) - f(0) - \frac{bx}{1-t}\right| \leqslant \frac{\varepsilon |x|}{1-t}.$$ After dividing through by $|x|$, you should be able to see that $f'(0)$ exists, and is equal to $\dfrac b{1-t}.$
 
  • #10
Opalg said:
ZaidAlyafey's comment shows that if $f'(0)$ exists then it is equal to $\dfrac b{1-t}.$ The snag with this, and with all the other comments, is that you are not told that $f$ is differentiable. You are not even told that it is continuous, except at the single point $0$. So you cannot use l'Hôpital's rule, or the mean value theorem, or any other result that requires differentiability, and you need to prove that $f$ is differentiable at $0$.

You are told that \lim_{x\rightarrow 0}\frac{f(x)-f(tx)}{x}=b. So given $\varepsilon>0$ there exists $\delta>0$ such that $$\left|\frac{f(x)-f(tx)}{x} - b\right| < \varepsilon$$ whenever $|x|<\delta.$ Therefore $$|f(x) - f(tx) - bx| < \varepsilon|x|.$$ Next, if $|x|<\delta$ then $|t^nx|<\delta$ for all $n\geqslant1$ (because $t<1$). So in the above inequality we can replace $x$ by $tx$, $t^2x,\ldots, t^{n-1}x,$ to get $$|f(tx) - f(t^2x) - btx| < \varepsilon t|x|,$$ $$|f(t^2x) - f(t^3x) - bt^2x| < \varepsilon t^2|x|,$$ $$\ldots$$ $$|f(t^{n-1}x) - f(t^nx) - bt^{n-1}x| < \varepsilon t^{n-1}|x|.$$ Now add those inequalities and use the triangle inequality, to get $$\left|\sum_{k=0}^{n-1}\bigl(f(t^kx) - f(t^{k+1}x) - bt^kx\bigr)\right| \leqslant \sum_{k=0}^{n-1}\bigl|f(t^kx) - f(t^{k+1}x) - bt^kx\bigr| < \sum_{k=0}^{n-1} \varepsilon t^k|x|.$$ It follows, by summing the geometric and telescoping series, that $$\left| f(x) - f(t^nx) - \frac{b(1-t^n)x}{1-t} \right| < \frac{\varepsilon(1-t^n)|x|}{1-t}.$$ Now you can let $n\to\infty$ and use the facts that $t^nx\to0$ and $f$ is continuous at $0$, to deduce that $$\left| f(x) - f(0) - \frac{bx}{1-t}\right| \leqslant \frac{\varepsilon |x|}{1-t}.$$ After dividing through by $|x|$, you should be able to see that $f'(0)$ exists, and is equal to $\dfrac b{1-t}.$

I understand!Thank you very much! :D
 

Similar threads

  • · Replies 8 ·
Replies
8
Views
2K
  • · Replies 23 ·
Replies
23
Views
3K
  • · Replies 11 ·
Replies
11
Views
2K
  • · Replies 0 ·
Replies
0
Views
2K
  • · Replies 9 ·
Replies
9
Views
2K
  • · Replies 4 ·
Replies
4
Views
3K
  • · Replies 22 ·
Replies
22
Views
2K
  • · Replies 11 ·
Replies
11
Views
2K
  • · Replies 1 ·
Replies
1
Views
2K
  • · Replies 24 ·
Replies
24
Views
4K