If we are given a complete metric space X and $f$ a contraction from into itself, then the Banach fixed Point theorem assures the existance of a unique fixed Point. How about the reverse of this conclusion, i.e. given a fixed point of a function defined on a complete metric space, can we suggest that it will be a contraction close to the fixed point ? If not, under which conditions it will be a contraction ?

  • $\begingroup$ No. Take a linear (or "affine") function $f:\mathbb{R}\rightarrow\mathbb{R}$ and play around with the slope. $\endgroup$ – Michael Oct 16 '17 at 22:16
  • $\begingroup$ A more interesting question would be the following: Given we have a unique fixed point $x_0$ of $f$ in $X$, is there a bijective mapping $g : X\to X$ such that $g^{-1}\circ f\circ g$ is a contraction and $g(x_0) = x_0$? $\endgroup$ – amsmath Oct 16 '17 at 22:59

Let $f:\Bbb R\to\Bbb R$, $f(x)=2x$. The unique fixed point is $0$, while $f$ is celarly not a contraction, rather expanding map.

  • $\begingroup$ Thanks. In the case it is not differentiable at the fixed point, i want to figure out under which conditions $f$ will be a contraction ? $\endgroup$ – user249018 Oct 16 '17 at 22:37
  • $\begingroup$ A contraction in $\mathbb R$ is a.e. differentiable since it is Lipschitz continuous and therefore absolutely continuous and therefore of bounded variation. $\endgroup$ – amsmath Oct 16 '17 at 22:47
  • $\begingroup$ A.e. differentiable means that it must not be differentiable at the fixed point. Do you agree ? $\endgroup$ – user249018 Oct 16 '17 at 22:59
  • $\begingroup$ @user249018 Completely wrong. There are many differentiable functions with fixed points. A differentiable function is (in particular) almost differentiable. $\endgroup$ – szw1710 Oct 17 '17 at 6:25

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.