0
$\begingroup$

Suppose $f(x)=x^i$ as a simple polynomial function of power $i$. As we know from the definition of the derivative operators, such an operator must be linear. How can we prove that the derivative operator of polynomials is actually a linear operator?
I am specifically interested in using Fréchet derivative formula to give a proof.

$\endgroup$
  • $\begingroup$ You can actually show that the derivative is linear for every function using the limit definition. $\endgroup$ – Javi Aug 11 '18 at 9:43
  • $\begingroup$ @Javi but the limit definition of derivative requires the assumption of the linearity of the operator; it can not be used to show linearity. $\endgroup$ – Mehdi Moghadasian Aug 11 '18 at 10:01
  • 1
    $\begingroup$ It only uses linearity of the limit, which can also prove if you dare. $\endgroup$ – Javi Aug 11 '18 at 10:10
  • $\begingroup$ @Javi Thanks for your response. Consider Fréchet derivative. If there exists a bounded linear operator $D(.): V\rightarrow U$ such that $\lim_{h\to 0}\frac{||f(x+h)-f(x)-Df(x)h||}{||h||}=0$ holds, then $D(.)$ is the unique derivative of $f(x)$. Do you mean if this operator exists, it is necessarily a linear operator? Or you are suggesting to show the linearity by using the definition? If latter is what you mean, how can I show the linearity principle $D(rf+sg)(x)=rDf(x)+sDg(x)$ by using the limit definition? Many thanks for your help. $\endgroup$ – Mehdi Moghadasian Aug 11 '18 at 10:56
  • 1
    $\begingroup$ Many thanks @Javi . Yes, I'd forgotten to mention Frechet. I edited my question. $\endgroup$ – Mehdi Moghadasian Aug 11 '18 at 14:35
1
$\begingroup$

\begin{align} &\lim_{h \to 0} \frac{|r(x+h)^i+s(x+h)^j-(rx^i+sx^j)-(irx^{i-1}-jsx^{j-1})h|}{|h|}\\ &=\lim_{h \to 0}\left|r\left(\frac{(x+h)^i-x^i}{h} -ix^{i-1}\right)+s\left(\frac{(x+h)^j-x^j}{h} -jx^{j-1}\right)\right|\\ &=\left|r\lim_{h \to 0}\left(\frac{(x+h)^i-x^i}{h} -ix^{i-1}\right)+s\lim_{h \to 0}\left(\frac{(x+h)^j-x^j}{h} -jx^{j-1}\right)\right|\\ &=0 \end{align}

where I have used continuity of norm and linearity of limit.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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