No label defined (Q5313262)

The paper is mainly concerned with the global behavior of the solutions of the third order rational difference equation \[ x_{n+1}=\frac{x_n+ \gamma x_{n-1}}{x_n+Dx_{n-2}},\quad n=0,1,2,\dots \tag{*} \] with nonnegative parameters, \(\gamma,~D\), and nonnegative initial conditions, \(x_{-2},~x_{-1},~x_0\). In this note, the authors (1) prove that if \(\gamma>1\), then Eq. (*) has an unbounded solution, and conjecture that every bounded solution of Eq. (*) converges to the unique equilibrium solution \(\bar{x}=\frac{1+\gamma}{1+D}\). (2) point out that if \(\gamma=1\), then all prime period-two solutions are of the form \(...,\phi,\psi,...\) with \(\phi,~\psi \in (0,\infty),~\phi\neq \psi,~\text{and}~ (1+D)\phi \psi=\phi+\psi\), and conjecture that every positive solution of Eq. (*) converges to a period-two solution of Eq. (*). (3) prove, by linearizing about the equilibrium solution \(\bar{x}\), that if \(\gamma<1\) and \(D\leq 1+\sqrt{2}\), or if \(D>1+\sqrt{2}\) and \(\frac{D^2-2D+1}{D^2+5D+2}<\gamma<1\), then \(\bar{x}\) is locally asymptotically stable, and conjecture that under these conditions, \(\bar{x}\) is globally asymptotically stable. Also, they conjecture that if \(\gamma<1\), then Eq. (*) has a bounded solution that does not converge to \(\bar{x}\). \noindent Furthermore, they prove that \begin{itemize}\item[-] if \(\gamma<D<1\), then the interval \([\frac{\gamma}{D},1]\) is an invariant interval for Eq. (*), and every solution of Eq. (*) with three consecutive terms in this interval converges to \(\bar{x}\); \item[-] if \(D<\gamma<1\), then the interval \([1,\frac{\gamma}{D}]\) is an invariant interval for Eq. (*), and every solution of Eq. (*) with three consecutive terms in this interval converges to \(\bar{x}\).\end{itemize}