New Reply

How to take this derivative?

 
Share Thread
Jul1-12, 03:15 PM   #1
 

How to take this derivative?


I promise guys, no homework here, just curiosity.

I am trying to find dy/dx for the equation y = x^x - C, where C is any arbitrary constant. I've found two ways that SHOULD be ok to take this derivative, but they produce different answers, I was wondering which method is correct and which method is incorrect. Also, why? It seems to me that both of these methods should be ok. Anyways, here they are:

Method 1:
[itex]y = x^x - C[/itex]
[itex]ln(y) = ln(x^x - C)[/itex]
[itex]ln(y) = \frac{x ln(x)}{ln(C)}[/itex]
Now take the derivative:
[itex]\frac{\frac{dy}{dx}}{y}=\frac{1}{ln(C)}(x ln(x))'[/itex]
Using the Product Rule, it can be seen that [itex](x ln(x))' = ln(x) + 1[/itex]. Therefore:
[itex]\frac{dy}{dx}=\frac{y}{ln(C)}(ln(x) + 1)[/itex]
[itex]\frac{dy}{dx}=\frac{x^x}{ln(C)}(ln(x) + 1)[/itex]

Method 2:
[itex]y = x^x - C[/itex]
[itex]y' = (x^x)' - C'[/itex]
[itex]y' = (x^x)'[/itex]
[itex](x^x)'[/itex] can be evaluated using method 1 for the equation [itex]y = x^x[/itex]
[itex]\frac{dy}{dx} = x^x(ln(x) + 1)[/itex]

Method one seems a bit less hand wavy, so I'm more confident in it; however, the derivative shouldn't depend on C, so that makes me lean more toward Method 2.

Anybody have any input they'd be willing to share?
PhysOrg.com mathematics news on PhysOrg.com

>> Pendulum swings back on 350-year-old mathematical mystery
>> Bayesian statistics theorem holds its own - but use with caution
>> Math technique de-clutters cancer-cell data, revealing tumor evolution, treatment leads
Jul1-12, 03:36 PM   #2
 
rewrite x as e^ln(x) to e^(xln(x)) and then use the rule d/dx (e^u) = e^u * u'
Jul1-12, 03:42 PM   #3
 
[itex]\ln(a-b)\neq\ln a/\ln b[/itex]. If I were you I'd just try to rewrite the [itex]x^x[/itex] as [itex]e^{f(x)}[/itex].
Jul1-12, 04:49 PM   #4
 

How to take this derivative?


Quote by jedishrfu View Post
rewrite x as e^ln(x) to e^(xln(x)) and then use the rule d/dx (e^u) = e^u * u'
Rewrite which x? Both? In the original or in the derivative?

Are you saying to do this?:

[itex]y = x^x - C[/itex]
[itex]e^y = e^{x^x - C} = e^{x^x}/e^C[/itex]

And then differentiate?

Quote by Ibix View Post
[itex]\ln(a-b)\neq\ln a/\ln b[/itex]. If I were you I'd just try to rewrite the [itex]x^x[/itex] as [itex]e^{f(x)}[/itex].
Oh right!! I was getting the rule backwards it seems.

So would method 2 be the one that is correct then?

y = x^x can be differentiated pretty easily with implicit differentiation.
Jul1-12, 05:02 PM   #5
 
Recognitions:
Homework Helper Homework Help
Quote by jedishrfu View Post
rewrite x as e^ln(x) to e^(xln(x)) and then use the rule d/dx (e^u) = e^u * u'
Quote by Hertz View Post
Rewrite which x? Both? In the original or in the derivative?

Are you saying to do this?:

[itex]y = x^x - C[/itex]
[itex]e^y = e^{x^x - C} = e^{x^x}/e^C[/itex]

And then differentiate?
jedishrfu is saying that
[itex]x = e^{\ln x}[/itex]
so
[itex]x^x = \left(e^{\ln x}\right)^x = e^{x\ln x}[/itex]

So what you need to take the derivative of is this:
[itex]y = e^{x\ln x} - C[/itex]
and you won't need to worry about implicit differentiation.
Jul1-12, 05:16 PM   #6
 
Quote by eumyang View Post
jedishrfu is saying that
[itex]x = e^{\ln x}[/itex]
so
[itex]x^x = \left(e^{\ln x}\right)^x = e^{x\ln x}[/itex]

So what you need to take the derivative of is this:
[itex]y = e^{x\ln x} - C[/itex]
and you won't need to worry about implicit differentiation.
Oh I see, well let's try:

[itex]y = e^{x\ln x} - C[/itex]
[itex]\frac{dy}{dx} = (e^{x\ln x})'[/itex]
[itex]\frac{dy}{dx} = e^{x\ln x}(ln(x) + 1)[/itex]
[itex]\frac{dy}{dx} = x^x(ln(x) + 1)[/itex]

Thanks for the help guys. I think the only problem I was having was mixing up my logarithm properties :)
Jul1-12, 05:47 PM   #7
 
Recognitions:
Gold Membership Gold Member
Science Advisor Science Advisor
Retired Staff Staff Emeritus
The derivative of an added or subtracted constant is 0 so the derivative of
[itex]y= x^x- C[/itex] is exactly the same as the derivative of [itex]y= x^x[/itex]. Now, y= x ln(x) so dy/dx= ln(x)- (x/x)= ln(x)- 1.
Jul1-12, 06:16 PM   #8
 
Recognitions:
Homework Helper Homework Help
Science Advisor Science Advisor
Quote by HallsofIvy View Post
The derivative of an added or subtracted constant is 0 so the derivative of
[itex]y= x^x- C[/itex] is exactly the same as the derivative of [itex]y= x^x[/itex]. Now, y= x ln(x) so dy/dx= ln(x)- (x/x)= ln(x)- 1.
I think you meant: ln(y)= x ln(x) so (dy/dx)/y = ln(x) + (x/x)= ln(x)+ 1 (Note: +, not -).
This leads to the same answer as in the OP by Method 2.
Method 1 went wrong after the second line. ln(A-B) is not ln(A)/ln(B). Must be thinking of ln(A/B) = ln(A) - ln(B) (or, equivalently, exp(A-B) = exp(A)/exp(B)).
New Reply

Similar discussions for: How to take this derivative?
Thread Forum Replies
Derivative with respect to partial derivative of contravariant metric tensor density Advanced Physics Homework 0
Find the derivative and the domain of the derivative (trig funtions) Calculus & Beyond Homework 7
find derivative of floor function using limit definition of derivative? Calculus & Beyond Homework 10
replacing total derivative with partial derivative in Griffiths' book Advanced Physics Homework 3
Intuitively what's the difference between Lie Derivative and Covariant Derivative? Differential Geometry 3