Register to reply

Derivative of the pi function

by 3.14159253589
Tags: derivative, function
Share this thread:
3.14159253589
#1
Aug6-11, 02:10 AM
P: 5
What is the derivative of the pi function
Phys.Org News Partner Science news on Phys.org
Apple to unveil 'iWatch' on September 9
NASA deep-space rocket, SLS, to launch in 2018
Study examines 13,000-year-old nanodiamonds from multiple locations across three continents
dimension10
#2
Aug6-11, 02:27 AM
P: 371
If you are talking about the prime number thing, I am not sure about the exact one but here is an approximation:

[tex]\pi(n)\approx \int_{2}^{n}\frac{dt}{\mbox{ln}(t)}[/tex]

[tex]\frac{d}{dn}\pi(n)\approx \frac{d}{dn}\int_{2}^{n}\frac{dt}{\mbox{ln}(t)}[/tex]

[tex]\frac{d}{dn}\pi(n)\approx \frac{d}{dn}\lim_{\delta t \rightarrow 0}\sum_{t=2}^{n}\frac{\delta t}{\mbox{ln}(t)}[/tex]

As the derivative of a sum is the sum of the derivatives,

[tex]\frac{d}{dn}\pi(n)\approx \lim_{\delta t \rightarrow 0}\sum_{t=2}^{n}\frac{d}{dn}\frac{\delta t}{\mbox{ln}(t)}[/tex]

[tex]\frac{d}{dn}\pi(n)\approx \lim_{\delta t \rightarrow 0}\sum_{t=2}^{n}-\frac{\delta t}{n \; {\mbox{ln}}^{2}(n)}[/tex]

[tex]\frac{d}{dn}\pi(n)\approx - \int_{2}^{n}\frac{dt}{n \; {\mbox{ln}}^{2}(n)} [/tex]

So that is the approximate rate of change of the pi function of t as t changes.
I like Serena
#3
Aug6-11, 10:55 AM
HW Helper
I like Serena's Avatar
P: 6,187
This is not quite right.
The proper derivative is:
[tex]\frac{d}{dn}\pi(n)\approx \frac{d}{dn}\int_{2}^{n}\frac{dt}{\ln(t)} = \frac{1}{\ln(n)}[/tex]

dimension10
#4
Aug6-11, 07:29 PM
P: 371
Derivative of the pi function

Quote Quote by I like Serena View Post
This is not quite right.
The proper derivative is:
[tex]\frac{d}{dn}\pi(n)\approx \frac{d}{dn}\int_{2}^{n}\frac{dt}{\ln(t)} = \frac{1}{\ln(n)}[/tex]
Then where did I make a mistake?
dimension10
#5
Aug6-11, 07:38 PM
P: 371
According to Wolfram Alpha,

[tex] \frac{d}{dn}(\int_{2}^{n}\frac{dt}{\ln(t)})=\frac{-n+n\; \mbox{ln}(n)+2}{n \;{\mbox{ln}}^{2}(n)}[/tex]

that is,

[tex] {\mbox{ln}}^{-1}(n)+(2{n}^{-1}{\mbox{ln}}^{-2}(n))-{\mbox{ln}}^{-2}(n)[/tex]

But for Li(n) alone, it gives [tex]{\mbox{ln}}^{-1}(n)[/tex] which is your solution.

And my solution gives [tex]- \int_{2}^{n}\frac{dt}{n \; {\mbox{ln}}^{2}(n)}[/tex] which is equal to [tex]\frac{n-2}{n\; {\mbox{ln}}^{2}(n)}[/tex]

I guess all three solutions are equal to each other and thus, correct?
MathematicalPhysicist
#6
Aug6-11, 11:06 PM
P: 3,243
Quote Quote by dimension10 View Post
If you are talking about the prime number thing, I am not sure about the exact one but here is an approximation:

[tex]\pi(n)\approx \int_{2}^{n}\frac{dt}{\mbox{ln}(t)}[/tex]

[tex]\frac{d}{dn}\pi(n)\approx \frac{d}{dn}\int_{2}^{n}\frac{dt}{\mbox{ln}(t)}[/tex]

[tex]\frac{d}{dn}\pi(n)\approx \frac{d}{dn}\lim_{\delta t \rightarrow 0}\sum_{t=2}^{n}\frac{\delta t}{\mbox{ln}(t)}[/tex]

As the derivative of a sum is the sum of the derivatives,

[tex]\frac{d}{dn}\pi(n)\approx \lim_{\delta t \rightarrow 0}\sum_{t=2}^{n}\frac{d}{dn}\frac{\delta t}{\mbox{ln}(t)}[/tex]

[tex]\frac{d}{dn}\pi(n)\approx \lim_{\delta t \rightarrow 0}\sum_{t=2}^{n}-\frac{\delta t}{n \; {\mbox{ln}}^{2}(n)}[/tex]

[tex]\frac{d}{dn}\pi(n)\approx - \int_{2}^{n}\frac{dt}{n \; {\mbox{ln}}^{2}(n)} [/tex]

So that is the approximate rate of change of the pi function of t as t changes.
I believe the problem is here:
[tex]\frac{d}{dn}\pi(n)\approx \lim_{\delta t \rightarrow 0}\sum_{t=2}^{n}\frac{d}{dn}\frac{\delta t}{\mbox{ln}(t)}[/tex]

The derivative operator [tex]\frac{d}{dn}[/tex] should operate on the sum.
I am not even sure what it means to have delta t inside a sum where t is a dumby variable, and then taking a limit of it as it approaches zero, the notation here is quite problematic.
I like Serena
#7
Aug7-11, 03:36 AM
HW Helper
I like Serena's Avatar
P: 6,187
Quote Quote by dimension10 View Post
Then where did I make a mistake?
As MathematicalPhysicist already said, the first mistake is when you moved d/dn to the other side of the summation symbol.
This is not allowed, because the summation is dependent on n.

You made another mistake when you differentiated the expression dependent on t with respect to n.
Since the expression is not dependent on n, the result is zero.


Quote Quote by dimension10 View Post
According to Wolfram Alpha,

[tex] \frac{d}{dn}(\int_{2}^{n}\frac{dt}{\ln(t)})=\frac{-n+n\; \mbox{ln}(n)+2}{n \;{\mbox{ln}}^{2}(n)}[/tex]
How did you get WolframAlpha to say that?
I do not get that.


Quote Quote by dimension10 View Post
I guess all three solutions are equal to each other and thus, correct?
The three solutions are not equal to each other, so they cannot all be correct.
dimension10
#8
Aug10-11, 04:51 AM
P: 371
Quote Quote by I like Serena View Post
As MathematicalPhysicist already said, the first mistake is when you moved d/dn to the other side of the summation symbol.
This is not allowed, because the summation is dependent on n.
Thanks. I forgot about that.

Quote Quote by I like Serena View Post
You made another mistake when you differentiated the expression dependent on t with respect to n.
Since the expression is not dependent on n, the result is zero.
Oops.


Quote Quote by I like Serena View Post
How did you get WolframAlpha to say that?
I do not get that.

I think I know what happened. It must have again considered d as constant rather than an infinitesimal.There seems to be a simpler solution using the second fundamental theorem of calculus and that would yield

[tex]\frac{1}{\mbox{ln}(n)}[/tex]

which is the answer given by you.
zetafunction
#9
Aug31-11, 10:28 AM
P: 399
the derivative of the prime counting function is just the sum

[tex] \delta (x-p) [/tex] taken over all primes 'p'
Robert1986
#10
Aug31-11, 07:21 PM
P: 828
I am confused; pi is a step function and is therefore only differentiable at the "interior" of a step at which point it is 0.
zetafunction
#11
Sep8-11, 10:32 AM
P: 399
yes but since the step function is discontinous delta function appear whenever the function has discontinuties, in the case of Pi function the discontinuities are located at the prime numbers


Register to reply

Related Discussions
Prove that derivative of the theta function is the dirac delta function Advanced Physics Homework 3
Find derivative of floor function using limit definition of derivative? Calculus & Beyond Homework 10
Function - derivative (help!) Calculus & Beyond Homework 2
First & Second derivative of a function Calculus & Beyond Homework 5
Derivative of function Calculus & Beyond Homework 5