Constructing means on certain continua and functional equations (Q791773): Difference between revisions

A mean on a topological space M is any continuous function \(\mu\) from \(M\times M\) onto M such that \(\mu(x,y)=\mu(y,x)\) and \(\mu(x,x)=x\) for all x,\(y\in M\). Fix a sequence \((f_ n:n\in {\mathbb{N}})\) of continuous self- mappings of the unit interval I and put \(M=\{(x_ 1,x_ 2,...)\in I^{{\mathbb{N}}}:f_ n(x_{n+1})=x_ n\quad \forall n\in {\mathbb{N}}\}\). If \((\mu_ n:n\in {\mathbb{N}})\) is a sequence of means on I such that \(f_ n(\mu_{n+1}(x,y))=\mu_ n(f_ n(x),f_ n(y))\) for all x,\(y\in I\), then the function \(\mu\) :\(M\times M\to M\) defined by the formula \(\mu((x_ 1,x_ 2,...),(y_ 1,y_ 2,...))=(\mu_ 1(x_ 1,y_ 1),\mu_ 2(x_ 2,y_ 2),...)\) is a mean on M and this mean is called an inverse limit mean generated by the sequence \((\mu_ n:n\in {\mathbb{N}})\). If M admits an inverse limit mean generated by a sequence \((\mu_ n:n\in {\mathbb{N}})\) such that for infinitely many n either \(\mu_ n=\max\), or \(\mu_ n=\min\), or \(0<x<y<1\) implies \(x<\mu_ n(x,y)<y\), then M is an arc. Put \[ g(x)= \begin{cases} 2x,\quad& x\in [0,1/2], \\ 2-2x,\quad& x\in(1/2,1] \end{cases}. \] It turns out that there exists no sequence of means \((\mu_ n:\) \(n\in {\mathbb{N}})\) such that \(g(\mu_{n+1}(x,y))=\mu_ n(g(x),g(y))\) for all x,\(y\in I\) and \(n\in {\mathbb{N}}\). Moreover, if J is a closed and bounded interval, h:\(J\to J\) is a conjugate of g (i.e., \(h=\Phi {\mathbb{O}}g{\mathbb{O}}\Phi^{-1}\), where \(\Phi\) :\(J\to^{onto}J\) is a homeomorphism) and \(\mu\) :\(J\times J\to J\) is a continuous function such that \(h(\mu(x,y))=\mu(h(x),h(y))\) for all x,\(y\in J\) and the function ''\(x\mapsto \mu(x,x)\), \(x\in J''\) is a bijection of J, then either \(\mu(x,y)=\mu(x,x)\) for all x,\(y\in J\) or \(\mu(x,y)=\mu(y,y)\) for all x,\(y\in J\).