3

Can you help me to proof that the nth derivative of $9\sqrt{x}$ is $$ (-1)^{(n-1)} \cdot \frac{9(2n-2)!}{(n-1)!} \cdot (4x)^{\frac{1-2n}{2}}$$

I've tried induction but didn't go very far.

Many thanks

3 Answers3

3

Firstly, the 9 is unimportant. Induction is the correct way to proceed, but perhaps we can disguise it in a computation:

$$\begin{align} \frac{d^n}{dx^n} \left( x^{\frac{1}{2}} \right) &= \frac{d^{n-1}}{dx^{n-1}} \left( \frac{1}{2} x^{\frac{1}{2} -1} \right) \\&= \frac{d^{n-2}}{dx^{n-2}}\left( \frac{1}{2} \left(\frac{1}{2} - 1 \right) x^{\frac{1}{2} -2} \right) \\&= \dots =\frac{d^{n-k}}{dx^{n-k}} \left( \frac{1}{2} \left( \frac{1}{2} -1\right) \left( \frac{1}{2} -2 \right)\dots \left(\frac{1}{2} -(k-1) \right) x^{\frac{1}{2} - k}\right) \\&= \frac{1}{2} \left( \frac{1}{2} - 1 \right) \left( \frac{1}{2} -2 \right)\dots \left(\frac{1}{2} - (n-1) \right) x^{\frac{1}{2} -n }. \end{align} $$ Can you simplify this?

3

Induction is a very good way. For the induction step,

Say $P(n)$ is true for $n=k$
For $P(k+1)$, $$\frac{d}{dx^{k+1}}(9\sqrt x)$$ $$=\frac{d}{dx}\left(\frac{d}{dx^{k}}(9\sqrt x)\right)$$ $$=\frac{d}{dx} \left[(-1)^{(k-1)} \cdot \frac{9(2k-2)!}{(k-1)!} \cdot (4x)^{\frac{1-2k}{2}}\right]$$ $$=(-1)^{(k-1)} \cdot \frac{9(2k-2)!}{(k-1)!} \cdot \frac{d}{dx} \left[(4x)^{\frac{1-2k}{2}}\right]$$ $$=(-1)^{(k-1)} \cdot \frac{9(2k-2)!}{(k-1)!} \cdot \frac{1-2k}{2} (4x)^{\frac{-1-2k}{2}}\cdot 4$$ $$=(-1)^{[(k+1)-1]} \cdot \frac{9(2k-2)!}{(k-1)!} \cdot (2k-1) (4x)^{\frac{-1-2k}{2}}\cdot 2$$ $$=(-1)^{[(k+1)-1]} \cdot \frac{9(2k-2)!}{(k-1)!} \cdot (2k-1) \cdot \frac{2k}{k} \cdot (4x)^{\frac{-1-2k}{2}}$$ $$=(-1)^{[(k+1)-1]} \cdot \frac{9[2(k+1)-2]!}{[(k+1)-1]!} \cdot (4x)^{\frac{1-2(k+1)}{2}}$$

$P(k+1)$ is true.

  • Induction is what makes more sense to me. I got to half of this but you got me through the place where I was stuck. – user290335 Nov 15 '15 at 17:42
  • You're welcome. – SchrodingersCat Nov 15 '15 at 17:43
  • In your 5th line you are multiplying by 4. How did you do that? If I split the power, as a product I will not get the same. – user290335 Nov 15 '15 at 18:27
  • I am not multiplying by $4$. I differentiated the entire expression first with respect to $4x$ and then differentiated $4x$ with respect to $x$ to get $4$. Get it? If you split the power as a product, you will get the same. Consider the exponent of $4$. If you split the power as a product, power of $4$ will remain same i.e. $\frac{1-2k}{2}$ and in my working, add the exponents of $4$, you will find it same. Clear?? – SchrodingersCat Nov 16 '15 at 04:53
  • Well I got what you did but still don't understand how splitting it into $4^{\frac{1-2k}{2}} . x^{\frac{1-2k}{2}}$ does not produce the same. Nevertheless, moving on, I do not understand how you go from $(1-2k )$ to $(2k-1)$ and the exponent of $(-1)$ goes from $(k-1)$ to $[(k+1)-1]$. Thanks – user290335 Nov 16 '15 at 15:10
  • For the first thing that you were talking of, it produces the same even on splitting the power as product. And for the second thing, I actually take $(-1)$ common from $(1-2k)$ and multiply it with $(-1)^{k-1}$ to get $(-1)^{k}$ and $(k+1)-1=k$. – SchrodingersCat Nov 16 '15 at 16:48
  • Hi, Can you also explain the last step where you multiply and divide by k, and how it affects the factorial? Thanks – user290335 Nov 18 '15 at 16:24
  • Do you mean the second last step? Thats where I have done so. – SchrodingersCat Nov 18 '15 at 16:25
  • Yes, which will produce the last equation. – user290335 Nov 18 '15 at 16:29
  • I simply multiply $2k(2k-1)$ with the numerator and $k$ with the denominator to get the respective factorials and then I have arranged them accordingly. Is it okay? – SchrodingersCat Nov 18 '15 at 16:31
2

Writing the first four derivatives of $f(x)=x^{1/2}$ reveals

$f^{(1)}(x)=\frac12 x^{-1/2}$

$f^{(2)}(x)=\frac12 \left(-\frac12\right)x^{-3/2}$

$f^{(3)}(x)=\frac12 \left(-\frac12\right)\left(-\frac32\right)x^{-5/2}$

$f^{(4)}(x)=\frac12 \left(-\frac12\right)\left(-\frac32\right)\left(-\frac52\right)x^{-7/2}$

We propose the general term is given by

$$f^{(n)}(x)=(-1)^{n-1}\left(\frac12\right)^n (2n-3)!!x^{-(2n-1)/2}$$

Then, the $n+1$ derivative

$$\begin{align} f^{(n+1)}(x)&=(-1)^{n-1}\left(\frac12\right)^{n}(-1)\left(\frac12\right) (2n-1)(2n-3)!!x^{-(2n-3)/2}\\\\ &=(-1)^n\left(\frac12\right)^{n+1}(2n-1)!!x^{-(2n-3)/2} \end{align}$$

which provides proof by induction!

Mark Viola
  • 179,405
  • 1
    This is really not a proof...the OP probably did this and deduced the formula. He/she is looking for a proof. – rogerl Nov 15 '15 at 17:39
  • @rogerl One can easily use induction using the expression of the general term, can one not? And why on earth did you single my answer out as others have posted similarly. – Mark Viola Nov 15 '15 at 17:41
  • Sure, but that's what the OP is having trouble with. – rogerl Nov 15 '15 at 17:41
  • @rogerl I added a bit more to provide the inductive step. Please let me know how I can improve my answer. I really want to give the best answer I can. – Mark Viola Nov 15 '15 at 17:51
  • Yes, I think that addresses the OPs question. – rogerl Nov 15 '15 at 19:00