Boundary value problems of discrete generalized Emden-Fowler equation (Q867785): Difference between revisions

This paper is concerned with the boundary value problem \[ \begin{aligned} \Delta \left( p_{k}\Delta u_{k-1}\right) +q_{k}u_{k} &= f(k,u_{k}),\;k=1,2,\dots ,n, \\ u_{0} &= A, \\ u_{n+1} &= B,\end{aligned} \] where \(p_{1},\dots,p_{n+1},q_{1},\dots,q_{n},A,B\) are real numbers such that each \( p_{i}\neq 0,\) and \(f:\left\{ 1,\dots ,n\right\} \times \mathbb{R}\rightarrow \mathbb{R}\) is continuous. It is easily established that \(u=(u_{0},u_{1},\dots ,u_{n+1})\) is a solution of this boundary value problem if, and only if, \( (u_{1},u_{2},\dots ,u_{n})\) is a solution of \(J^{\prime }(u)=0,\) where \(J(v)= \frac{1}{2}\left( Mv,v\right) +\left( \eta ,v\right) -F(v)\) for \(v\in \mathbb{R}^{n}\) and \[ M=\left(\begin{matrix} q_{1}-p_{1}-p_{2} & p_{2} & \cdots & 0 & 0 \\ p_{2} & q_{2}-p_{2}-p_{3} & \cdots & 0 & 0 \\ \cdots & p_{3} & \cdots & 0 & 0 \\ \cdots & \cdots & \cdots & \cdots & \cdots \\ \cdots & \cdots & \cdots & q_{n-1}-p_{n-1}-p_{n} & p_{n} \\ 0 & \cdots & \cdots & p_{n} & q_{n}-p_{n}-p_{n+1} \end{matrix} \right) , \] \[ \eta =\left( p_{1}A,0,\dots ,0,p_{n+1}B\right) ^{\dagger } \] and \[ F(v)=\sum_{j=1}^{n}\int_{0}^{v_{j}}f(j,s)ds. \] The function \(f\) is assumed to satisfy one of the following conditions: {\parindent=7mm \begin{itemize}\item[(i)]There are \(a_{1}>0,a_{2}>0,R>0,\alpha \in (1,2)\) such that \(F(v)\leq a_{1}\left\| v\right\| ^{\alpha }+a_{2}\) for all \(v\in \mathbb R^{n}\) that satisfies \(\left\| v\right\| >R.\) \item[(ii)]There is \(c_{0}>0\) such that \(\left| f(k,z)\right| \leq c_{0}\) for any \( (k,z)\in \{1,\dots ,n\}\times \mathbb{R}.\) \item[(iii)]There are \(a_{1}>0,a_{2}>0,R>0\) and \(\beta >2\) such that \(F(v)\geq a_{1}\left\| v\right\| ^{\beta }-a_{2}\) for all \(v\in \mathbb{R}^{n}\) that satisfies \( \left\| v\right\| \geq R.\) \item[(iv)]There is \(L>0\) such that \(\left| f(k,x)-f(k,y)\right| \leq L\left| x-y\right| \) for \(t\in \{1,\dots ,n\}\) and any \(x,y\in \mathbb{R}.\) \end{itemize}}Under additional conditions on the matrix \(M,\) existence criteria for solutions of critical points of \(J\) are established by means of a saddle point theorem. Some of the results in this paper have contact with those in the following references of the reviewer: \textit{S. S. Cheng} [J. Math. Anal. Appl. 111, 465--474 (1985; Zbl 0589.39002); Linear Algebra Appl. 85, 57--73 (1987; Zbl 0605.26017)]; \textit{S. S. Cheng} and \textit{S. S. Lin} [Util. Math. 39, 167--186 (1991; Zbl 0729.39002)]; \textit{S. S. Cheng} and \textit{R. F. Lu} [Fasc. Math. 23, 9--24 (1991; Zbl 0751.39002)].