• Support PF! Buy your school textbooks, materials and every day products via PF Here!

A limit problem without the use of a Taylor series expansion

Homework Statement
This is a famous problem of limit.
Lim ((1+x)^(1/x)-e+ex/2)/x^2 where x tends to 0.

We can easily solve it with taylor expansion but i want to solve it without using taylor expansion.Is there a way?
Homework Equations
Taylor expansion
I tried substituting x=cos2theta but it was of no use.I thought many ways but i could not make 0/0 form.So please help.
 

fresh_42

Mentor
Insights Author
2018 Award
11,421
7,842
Can you type this in LaTeX so that the function can be read more easily? See https://www.physicsforums.com/help/latexhelp/ for how it is done. I have read ##\lim_{x\to 0} \dfrac{(1+x)^{\frac{1}{x}}-e-e\cdot \frac{x}{2}}{x^2}## but I'm not sure. Have you tried L'Hôpital?
 
Can you type this in LaTeX so that the function can be read more easily? See https://www.physicsforums.com/help/latexhelp/ for how it is done. I have read ##\lim_{x\to 0} \dfrac{(1+x)^{\frac{1}{x}}-e-e\cdot \frac{x}{2}}{x^2}## but I'm not sure. Have you tried L'Hôpital?
Yes you have taken it correct.I think we cannot apply L' Hospital since 1^infinity is there and therefore no 0/0 or infinity/infinity form.
 

fresh_42

Mentor
Insights Author
2018 Award
11,421
7,842
And why isn't it simply ##-\infty##?
##\lim_{x\to 0} \dfrac{(1+x)^{\frac{1}{x}}-e-e\cdot \frac{x}{2}}{x^2}=
\lim_{x\to 0} \dfrac{e-e-e\cdot \frac{x}{2}}{x^2} =
\lim_{x\to 0} \dfrac{-e}{2x}##
 
And why isn't it simply ##-\infty##?
##\lim_{x\to 0} \dfrac{(1+x)^{\frac{1}{x}}-e-e\cdot \frac{x}{2}}{x^2}=
\lim_{x\to 0} \dfrac{e-e-e\cdot \frac{x}{2}}{x^2} =
\lim_{x\to 0} \dfrac{-e}{2x}##
So will L'Hospital work here.
 

fresh_42

Mentor
Insights Author
2018 Award
11,421
7,842
I haven't used L'Hôpital, only that ##\lim_{x \to 0} (x+1)^{\frac{1}{x}} =e##. I think that is a bit cheating, as I ignored the denominator, so it has to be shown formally that ##\lim_{x\to 0} \dfrac{(x+1)^{\frac{1}{x}}-e}{x^2}=0## by whatever method.
 
I tried L'Hospital but it is becoming complicated.
 
I haven't used L'Hôpital, only that ##\lim_{x \to 0} (x+1)^{\frac{1}{x}} =e##. I think that is a bit cheating, as I ignored the denominator, so it has to be shown formally that ##\lim_{x\to 0} \dfrac{(x+1)^{\frac{1}{x}}-e}{x^2}=0## by whatever method.
The answer is -11e/24.
 

fresh_42

Mentor
Insights Author
2018 Award
11,421
7,842

fresh_42

Mentor
Insights Author
2018 Award
11,421
7,842
This is because I made a sign error in post #2 which you confirmed in post #3. Please use LaTeX next time. Seems we cannot simplify ##(1+x)^{1/x}=e## in the limit. I had such a feeling. In this case I would try L'Hôpital, but Taylor is easier.
 
This is because I made a sign error in post #2 which you confirmed in post #3. Please use LaTeX next time. Seems we cannot simplify ##(1+x)^{1/x}=e## in the limit. I had such a feeling. In this case I would try L'Hôpital, but Taylor is easier.
Ok i will use LaTeX next time.But can you please solve it without Taylor.
 

fresh_42

Mentor
Insights Author
2018 Award
11,421
7,842
Ok i will use LaTeX next time.But can you please solve it without Taylor.
This means in other words: Can you prove that the third term of the Taylor series of ##\dfrac{(1+x)^\frac{1}{x}}{x^2}## equals ##\frac{11}{24}\,e## without using the Taylor series? However, already ##f''(0)## is an ugly limit. Another wording is: How fast exactly does ##(1+x)^\frac{1}{x}## converge to ##e##.

I'm sure there is a nice trick to manage those expressions, but I don't know any.
 
This means in other words: Can you prove that the third term of the Taylor series of ##\dfrac{(1+x)^\frac{1}{x}}{x^2}## equals ##\frac{11}{24}\,e## without using the Taylor series? However, already ##f''(0)## is an ugly limit. Another wording is: How fast exactly does ##(1+x)^\frac{1}{x}## converge to ##e##.

I'm sure there is a nice trick to manage those expressions, but I don't know any.
I am unable to converge that please help.
 

statdad

Homework Helper
1,492
32
The limit is not -11e/24 -- the limit does not exist.

a) You cannot expand the numerator in a series around 0 since the exponential term is not defined for x = 0, so that approach does not apply

b) More directly, a plot of the expression (look at the plot from (say) x = 0.001 to about x = 0.3) shows a vertical asymptote tending to negative infinity as you approach 0 from the right.
 

fresh_42

Mentor
Insights Author
2018 Award
11,421
7,842
The limit is not -11e/24 -- the limit does not exist.

a) You cannot expand the numerator in a series around 0 since the exponential term is not defined for x = 0, so that approach does not apply

b) More directly, a plot of the expression (look at the plot from (say) x = 0.001 to about x = 0.3) shows a vertical asymptote tending to negative infinity as you approach 0 from the right.
WolframAlpha says differently:
 

statdad

Homework Helper
1,492
32
I find Alpha's "series expansion" suspicious. The requirement for that to be valid are that the function in question be differentiable about the point in question (0, in this case), and this function is not defined at 0.

Also: it seems that there are two expressions being bantered about here. The original expression was

##
\lim_{x \to 0} \dfrac{\left(1+x\right)^{(1/x)} - e + \frac{ex}2}{x^2}
##

The other is
##
\lim_{x \to 0} \dfrac{\left(1+x\right)^{(1/x)} - e - \frac{ex}2}{x^2}
##

Which one is correct?
 

fresh_42

Mentor
Insights Author
2018 Award
11,421
7,842
I find Alpha's "series expansion" suspicious. The requirement for that to be valid are that the function in question be differentiable about the point in question (0, in this case), and this function is not defined at 0.

Also: it seems that there are two expressions being bantered about here. The original expression was

##
\lim_{x \to 0} \dfrac{\left(1+x\right)^{(1/x)} - e + \frac{ex}2}{x^2}
##

The other is
##
\lim_{x \to 0} \dfrac{\left(1+x\right)^{(1/x)} - e - \frac{ex}2}{x^2}
##

Which one is correct?
See post #11.
 

fresh_42

Mentor
Insights Author
2018 Award
11,421
7,842
No, I have no idea. I tried a few things but they led nowhere. The margin is very, very narrow and one will need many terms of the Taylor expansion of ##(1+x)^{1/x}##. I substituted ##n= 1/x## and worked with natural numbers, but calculation with just ##O(\frac{1}{n})## was not precise enough. Even integration uses the series expansion.
 
Last edited:

SammyS

Staff Emeritus
Science Advisor
Homework Helper
Gold Member
11,159
915
I find Alpha's "series expansion" suspicious. The requirement for that to be valid are that the function in question be differentiable about the point in question (0, in this case), and this function is not defined at 0.

Also: it seems that there are two expressions being bantered about here. The original expression was

##
\lim_{x \to 0} \dfrac{\left(1+x\right)^{(1/x)} - e + \frac{ex}2}{x^2}
##

The other is
##
\lim_{x \to 0} \dfrac{\left(1+x\right)^{(1/x)} - e - \frac{ex}2}{x^2}
##

Which one is correct?
The correct expression is ##\ \displaystyle
\lim_{x \to 0} \dfrac{\left(1+x\right)^{(1/x)} - e + \dfrac{e \cdot x}2}{x^2}
## .

Although the function, ## \left(1+x\right)^{(1/x)} ## is not defined at ## x=0 ##, this is a removable discontinuity.

The following function is defined and differentiable at ##x=0##, and has the Taylor expansion given by WolframAlpha for the function, ## \left(1+x\right)^{(1/x)} ##.

##\displaystyle \quad \quad u(x) = \begin{cases} \left(1+x\right)^{(1/x)} & \text{if } x \ne 0 \\ e & \text{if } x = 0 \end{cases} ##

.
 
The correct expression is ##\ \displaystyle
\lim_{x \to 0} \dfrac{\left(1+x\right)^{(1/x)} - e + \dfrac{e \cdot x}2}{x^2}
## .

Although the function, ## \left(1+x\right)^{(1/x)} ## is not defined at ## x=0 ##, this is a removable discontinuity.

The following function is defined and differentiable at ##x=0##, and has the Taylor expansion given by WolframAlpha for the function, ## \left(1+x\right)^{(1/x)} ##.

##\displaystyle \quad \quad u(x) = \begin{cases} \left(1+x\right)^{(1/x)} & \text{if } x \ne 0 \\ e & \text{if } x = 0 \end{cases} ##

.
Do you have any idea to solve it without taylor expansion.
 

SammyS

Staff Emeritus
Science Advisor
Homework Helper
Gold Member
11,159
915
Do you have any idea to solve it without Taylor expansion.
I don't have any good idea for that, but I'm thinking ...

As has been pointed out, the derivatives of the numerator get very complicated.
 

vela

Staff Emeritus
Science Advisor
Homework Helper
Education Advisor
14,411
1,099
Did you try L'Hopital's rule?
 

Want to reply to this thread?

"A limit problem without the use of a Taylor series expansion" You must log in or register to reply here.

Related Threads for: A limit problem without the use of a Taylor series expansion

Replies
11
Views
915
Replies
1
Views
529
Replies
8
Views
2K
Replies
1
Views
2K
Replies
5
Views
3K
  • Posted
Replies
5
Views
473
Replies
3
Views
869

Physics Forums Values

We Value Quality
• Topics based on mainstream science
• Proper English grammar and spelling
We Value Civility
• Positive and compassionate attitudes
• Patience while debating
We Value Productivity
• Disciplined to remain on-topic
• Recognition of own weaknesses
• Solo and co-op problem solving
Top