Theorem 9.24: Suppose $\textbf{f}$ is a $C'-$ mapping of an open set $E \subset \mathbb{R^n}$ into $\mathbb{R^n}$.$\textbf{f} \thinspace^\prime\textbf{ (a)}$ is invertible for some $\textbf{a}\in E$ and $\textbf{b=f(a)}$ then

(a) there exist open sets $U$ and $V$ in $\mathbb{R^n }$ such that $\textbf{a}\in U, \textbf{b}\in V,\textbf{f}$ is one-to one on $U$ and $\textbf{f}(U)=V$

(b) if $\textbf{g}$ is the inverse of $\textbf{f}$[which exist , by (a)], definedin $V$ by $\textbf{g(f(x))=x }$$\space $ $(\textbf{x}\in U)$ then $\textbf{g}\in C'(V)$

To prove (a) Rudin took a function $\varphi$ which is defined as to each $\textbf{y}\in \mathbb{R^n },\varphi(\textbf{x})=\textbf{x}+A^{-1}(\textbf{y}\thinspace \textbf{-f(x))}\space \space (\textbf{x}\in E )$(Equation 48) Then he states that $\varphi '(\textbf{x})=I- A^{-1}\textbf{f}\space'(x)$.I didn't got how rudin found $\varphi '(\textbf{x})$.Please help me to understand.

  • 1
    $\begingroup$ That looks like a basic calculus computation. What is the source of your doubt? $\endgroup$
    – Umberto P.
    Jul 4, 2018 at 17:21
  • $\begingroup$ Sorry I am not understanding this by basic calculus computation.Please help me $\endgroup$
    –  Mathan
    Jul 4, 2018 at 17:29

1 Answer 1


Note that $\varphi(x)$ is the sum of $x$ with $A^{-1}\bigl(y-f(x)\bigr)$. Well, $x$ is just the identity function, and $A^{-1}$ is linear. Finally, $y$ is constant. Therefore, by the rule for differentiating sums and by the chain rule (together with that fact that $(A^{-1})'=A^{-1}$), we get that\begin{align}\varphi'(x)&=\operatorname{Id}+A^{-1}\circ(y-f)'(x)\\&=\operatorname{Id}+A^{-1}\bigl(-f'(x)\bigr)\\&=\operatorname{Id}-A^{-1}\bigl(f(x)\bigr).\end{align}


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.