Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

I have a question:

Consider a function $g:\mathbb{R}^{n}\rightarrow \mathbb{R}$ is differentiable. Find the derivative of the function: $G(x)=[ g\left ( x,x^{2},...,x^{n} \right )]^{2}$ where $x\in \mathbb{R}$.

Here, G is a function of one variable, so I tried to apply the chain rule to find its derivative as follows: $G^{'}\left ( x \right )=2H^{'}\left ( x \right )H(x)$ where $H\left ( u_{1},u_{2},...,u_{n} \right )=g\left ( x,x^{2},...,x^{n} \right )$. Now to find the derivative of $H$, I did the following: $$\frac{d}{dx}H=\frac{\partial H}{\partial u_{1}}\frac{\partial u_{1}}{\partial x}+...\frac{\partial H}{\partial u_{n}}\frac{\partial u_{n}}{\partial x}.$$ Does what I did make sense?

share|cite|improve this question
Do you really mean to have the $\lfloor \;\;\rfloor$ (the floor function) in the definition of $G$? Perhaps you intended to have brackets $[\;\;]$? – Zev Chonoles Jan 29 '12 at 19:18
Provided one replaces $H(u_1,u_2,\ldots,u_n)$ by $H(x)$ and each $\partial H/\partial u_k$ by $\partial g/\partial u_k$, the answer is: yes. – Did Jan 29 '12 at 19:22
@ Zev Chonoles: Yes, I just mean simple brackets only. – pin296 Jan 29 '12 at 19:24
@pin296: Ah, I see that you actually did have brackets, but my browser was for some reason not displaying them correctly - apologies. – Zev Chonoles Jan 29 '12 at 19:28
Here is my answer based on the above comments: $G^{'}\left ( x \right )=2.g\left ( x,x^{2},...,x^{n} \right ).\left [ \frac{\partial g}{\partial u_{1}}.1+\frac{\partial g}{\partial u_{2}}.2x +...+nx^{n-1}.\frac{\partial g}{\partial u_{n}} \right ]$. where : $u_{i}=x^{i}$. I am not convinced to the presence of the variables $u_{i}=x^{i}$ in my final answer. Does anyone have a better idea on how to solve the problem? – pin296 Jan 29 '12 at 21:01

It's almost correct. You have a function $$u(\cdot):\quad {\mathbb R}\to{\mathbb R}^n\ ,\qquad x\mapsto u(x):=(x,x^2,\ldots, x^n)$$ and a second function $$g:\quad {\mathbb R}^n\to {\mathbb R}\ ,\qquad (u_1,\ldots, u_n)\mapsto g(u_1,\ldots, u_n)\ .$$ The real-valued function $H:x\mapsto H(x)$ is defined as the composition of the two: $$H(x)\ :=\ g\bigl(u(x)\bigr)\ .$$ By the chain rule the derivative of $H$ is given by $$H'(x)={\partial g\over\partial u_1} u_1'(x) +\ldots+ {\partial g\over\partial u_n} u_n'(x) =\nabla g\bigl(u(x)\bigr)\cdot u'(x)\ ,$$ where the $\cdot$ denotes the scalar product.

share|cite|improve this answer
I think my final answer looks like yours: $G^{'}\left ( x \right )=2.g\left ( x,x^{2},...,x^{n} \right ).\left [ \frac{\partial g}{\partial u_{1}}.1+\frac{\partial g}{\partial u_{2}}.2x +...+nx^{n-1}.\frac{\partial g}{\partial u_{n}} \right ]$. Is my expression true? – pin296 Jan 29 '12 at 21:06
Your "final answer" is correct. You can write $g_k$ instead of ${\partial g\over\partial u_k}$ if you want to avoid the appearance of the letter $u$ "out of nowhere". – Christian Blatter Jan 30 '12 at 8:52

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.