Let $f: \mathbb{R}^n \times \mathbb{R}^n \rightarrow \mathbb{R}^n$ and $g:\mathbb{R}^n \rightarrow \mathbb{R}^n$ be continuous and compact valued.

Consider the function $F: \mathbb{R}^n \times \mathbb{R}^n \rightarrow \mathbb{R}^n \times \mathbb{R}^n$ defined as $$ F(x,y) := \left[ \begin{matrix} f(x,y) \\ f(x,y) + g(y) + y \end{matrix} \right]. $$

Find conditions on $g$ such that $F$ has a fixed point.

Comments. I tried assuming that the function $y \mapsto g(y) + y$ has a fixed point $\bar{y}$, i.e., $\bar{y} = g( \bar y) + \bar y \Leftrightarrow g(\bar y) = 0$. But it does not seem to work.

  • $\begingroup$ what does compact valued mean? $\endgroup$ – user66081 Feb 5 '15 at 15:32
  • $\begingroup$ It means that the range of $f$ is a compact set. $\endgroup$ – user693 Feb 5 '15 at 16:57

At a fixed point of $F$ we must have $$\begin{bmatrix}x\\y\end{bmatrix}=F(x,y)=\begin{bmatrix}f(x,y)\\f(x,y)+g(y)+y\end{bmatrix}$$

i.e. $f(x,y)=x$ and $f(x,y)+g(y)+y=y$.

Therefore the system $$\begin{cases}f(x,y)&=x\\g(y)&=-x\end{cases}$$ must have solution.

In other words $f(-g(y),y)=-g(y)$ must have a solution. Conversely, if $f(-g(y),y)=-g(y)$ has a solution then $\begin{bmatrix}-g(y)\\y\end{bmatrix}$ is afixed point of $F$.

  • $\begingroup$ In the second statement, do you mean "if $f(-g(y),y)=-g(y)$" has solution? $\endgroup$ – user693 Feb 5 '15 at 16:47
  • $\begingroup$ @Adam Yes, the same equation. $\endgroup$ – Pp.. Feb 5 '15 at 16:51
  • $\begingroup$ Thanks. So your statement is: $F$ has a fixed point if and only if $y \mapsto f( -g(y), y ) + g(y)$ has a zero. $\endgroup$ – user693 Feb 5 '15 at 16:54
  • $\begingroup$ @Adam Yes, you can write it that way. $\endgroup$ – Pp.. Feb 5 '15 at 16:55
  • $\begingroup$ A follow up problem on the existence of such zero is in [math.stackexchange.com/questions/1142019/existence-of-a-zero]. Thanks! $\endgroup$ – user693 Feb 10 '15 at 12:58

The condition $F(x, y) = (x, y)$ implies $f(x, y) = x$, so $x = -g(y)$, so $f(-g(y), y) = -g(y)$.

Suppose $g$ is invertible with a continuous inverse, such that the range of $-f$ is contained in the range of $g$.

Consider $\psi : y \mapsto g^{-1}(-f(-g(y), y))$. This is a well defined, continuous function.

I speculate that one can to argue that $\psi(y) = y$ has a solution using the intermediate value theorem and boundedness of $f$. Then take $x := -g(y)$.

  • $\begingroup$ I guess $\psi(y) = y$ has solution because $\psi$ is continuous and compact valued, therefore it has a fixed point. $\endgroup$ – user693 Feb 5 '15 at 16:46
  • $\begingroup$ @Adam: well is it?, I don't know $\endgroup$ – user66081 Feb 5 '15 at 17:13

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.