Difference between revisions of "009C Sample Midterm 3, Problem 1"

From Grad Wiki
Jump to navigation Jump to search
(Created page with "<span class="exam">Test if the following sequence <math style="vertical-align: -10%">{a_n}</math> converges or diverges. If it converges, also find the limit of the sequence....")
 
m
Line 6: Line 6:
 
! Foundations: &nbsp;  
 
! Foundations: &nbsp;  
 
|-
 
|-
|This a common question, and is relates to the fact that
+
|This a common question, and is related to the fact that
 
|-
 
|-
 
|
 
|
 
::<math style="vertical-align: 0%">\lim_{x\rightarrow\infty}\left(1+\frac{\alpha}{x}\right)^{x}\ =\ e^{\alpha}.</math>  
 
::<math style="vertical-align: 0%">\lim_{x\rightarrow\infty}\left(1+\frac{\alpha}{x}\right)^{x}\ =\ e^{\alpha}.</math>  
 
|-
 
|-
|In such a limit, the argument <math style="vertical-align: -22%">1+\alpha /x</math> is always bigger than one, and tends to one as <math style="vertical-align: 0%">x</math> gets large, while we are raising that argument to an increasing power. Neither one really "wins", so we end up with a finite limit that is neither zero nor infinity.
+
|In such a limit, the argument <math style="vertical-align: -22%">1+\alpha /x</math> tends to one as <math style="vertical-align: 0%">x</math> gets large, while we are raising that argument to an increasing power. Neither one really "wins", so we end up with a finite limit that is neither zero nor infinity.
 
|-
 
|-
 
|On the other hand, in the exam problem the argument <math style="vertical-align: -22%">(n-7)/n</math> is always smaller than one, but tends to one as <math style="vertical-align: 0%">n</math> gets large, while the exponent <math style="vertical-align: -25%">1/n</math> tends to zero. These do not disagree, so the limit should be one, but we need to prove it.
 
|On the other hand, in the exam problem the argument <math style="vertical-align: -22%">(n-7)/n</math> is always smaller than one, but tends to one as <math style="vertical-align: 0%">n</math> gets large, while the exponent <math style="vertical-align: -25%">1/n</math> tends to zero. These do not disagree, so the limit should be one, but we need to prove it.
Line 44: Line 44:
 
&thinsp;so we can apply l'H&ocirc;pital's rule. Finally, since <math style="vertical-align: -60%">\ln L=-1,\,\,L=\frac{1}{e}.</math>  
 
&thinsp;so we can apply l'H&ocirc;pital's rule. Finally, since <math style="vertical-align: -60%">\ln L=-1,\,\,L=\frac{1}{e}.</math>  
 
|-
 
|-
|Again, such a technique is not required for this particular problem, as the exponent tends to zero.
+
|Again, such a technique is not required for this particular problem, as the exponent tends to zero.  But the technique is common enough on exams to justify providing an example.
 
|}
 
|}
  

Revision as of 11:30, 26 April 2015

Test if the following sequence converges or diverges. If it converges, also find the limit of the sequence.

Foundations:  
This a common question, and is related to the fact that
In such a limit, the argument Failed to parse (Conversion error. Server ("https://wikimedia.org/api/rest_") reported: "Cannot get mml. Server problem."): {\displaystyle 1+\alpha /x} tends to one as gets large, while we are raising that argument to an increasing power. Neither one really "wins", so we end up with a finite limit that is neither zero nor infinity.
On the other hand, in the exam problem the argument Failed to parse (Conversion error. Server ("https://wikimedia.org/api/rest_") reported: "Cannot get mml. Server problem."): {\displaystyle (n-7)/n} is always smaller than one, but tends to one as gets large, while the exponent Failed to parse (Conversion error. Server ("https://wikimedia.org/api/rest_") reported: "Cannot get mml. Server problem."): {\displaystyle 1/n} tends to zero. These do not disagree, so the limit should be one, but we need to prove it.
Any time you have a function raised to a function, we need to use natural log and take advantage of the log rule:
Failed to parse (Conversion error. Server ("https://wikimedia.org/api/rest_") reported: "Cannot get mml. Server problem."): {\displaystyle \ln \left(a^{b}\right)=b\ln(a).}
For example, to find Failed to parse (Conversion error. Server ("https://wikimedia.org/api/rest_") reported: "Cannot get mml. Server problem."): {\displaystyle \lim _{x\rightarrow \infty }\left(1-{\frac {1}{x}}\right)^{x}} , you could begin by saying: Let

Then

Failed to parse (Conversion error. Server ("https://wikimedia.org/api/rest_") reported: "Cannot get mml. Server problem."): {\displaystyle \ln L=\ln \left[\lim _{x\rightarrow \infty }\left(1-{\frac {1}{x}}\right)^{x}\right]=\lim _{x\rightarrow \infty }\ln \left[\left(1-{\frac {1}{x}}\right)^{x}\right],}

where we are allowed to pass the log through the limit because natural log is continuous. But by log rules,

Failed to parse (Conversion error. Server ("https://wikimedia.org/api/rest_") reported: "Cannot get mml. Server problem."): {\displaystyle \ln \left[\left(1-{\frac {1}{x}}\right)^{x}\right]=x\ln \left(1-{\frac {1}{x}}\right).}
Thus
Failed to parse (Conversion error. Server ("https://wikimedia.org/api/rest_") reported: "Cannot get mml. Server problem."): {\displaystyle {\begin{array}{rcl}\lim _{x\rightarrow \infty }\ln \left[\left(1-{\frac {1}{x}}\right)^{x}\right]&=&\lim _{x\rightarrow \infty }x\ln \left(1-{\frac {1}{x}}\right),\\&=&\lim _{x\rightarrow \infty }{\frac {\ln \left(1-{\frac {1}{x}}\right)}{\frac {1}{x}}}\\&{\overset {l'H}{=}}&\lim _{x\rightarrow \infty }{\frac {{\frac {x}{x-1}}\cdot \left(-{\frac {1}{x}}\right)'}{\left({\frac {1}{x}}\right)'}}\\&=&-1.\end{array}}}

Note that Failed to parse (Conversion error. Server ("https://wikimedia.org/api/rest_") reported: "Cannot get mml. Server problem."): {\displaystyle \lim _{x\rightarrow \infty }{\frac {\ln \left(1-{\frac {1}{x}}\right)}{\frac {1}{x}}}={\frac {0}{0}},}  so we can apply l'Hôpital's rule. Finally, since Failed to parse (Conversion error. Server ("https://wikimedia.org/api/rest_") reported: "Cannot get mml. Server problem."): {\displaystyle \ln L=-1,\,\,L={\frac {1}{e}}.}

Again, such a technique is not required for this particular problem, as the exponent tends to zero. But the technique is common enough on exams to justify providing an example.
Solution: 
Following the procedure outlined in Foundations, let Failed to parse (Conversion error. Server ("https://wikimedia.org/api/rest_") reported: "Cannot get mml. Server problem."): {\displaystyle L=\lim _{n\rightarrow \infty }\left({\frac {n-7}{n}}\right)^{1/n}.} Then
Failed to parse (Conversion error. Server ("https://wikimedia.org/api/rest_") reported: "Cannot get mml. Server problem."): {\displaystyle {\begin{array}{rcl}\ln L&=&\ln \left(\lim _{n\rightarrow \infty }\left[\left({\frac {n-7}{n}}\right)^{1/n}\right]\right)\\&=&\lim _{n\rightarrow \infty }\ln \left[\left({\frac {n-7}{n}}\right)^{1/n}\right]\\&&\lim _{n\rightarrow \infty }\left[{\frac {1}{n}}\cdot \ln \left({\frac {n-7}{n}}\right)\right]\\&=&0\cdot \ln(1)\\&=&0.\end{array}}}
Thus, Also, most teachers would require you to mention that natural log is continuous as justification for passing the limit through it.
Final Answer:  
The limit of the sequence is Failed to parse (Conversion error. Server ("https://wikimedia.org/api/rest_") reported: "Cannot get mml. Server problem."): {\displaystyle e^{0}=1.}

Return to Sample Exam