On nonlinear discrete boundary value problems (Q1823405)

The author considers two classes of boundary value problems for discrete problems of the form \[ (1\lambda)\quad y(t+1)=A(t)x(t)+h(t)+\lambda f(t,y(t)),\quad \sum^{N}_{j=0}B_ j(t_ j)=v+\lambda g(x(0),...,x(n)) \] and (2) \(x(t+1)=f(t,x(t))\), \(g(x(0),...,x(n))=0\), where \(t\in \{0,1,...,N\}\) is a fixed positive integer, \(B_ 0,...,B_ N\) are \(n\times n\) constant matrices, A(t) is an \(n\times n\) matrix and \(h(t)\in {\mathbb{R}}^ n\) for each t, \(v\in {\mathbb{R}}^ n\), \(\lambda\in {\mathbb{R}}\), and f: \({\mathbb{R}}^{n+1}\to {\mathbb{R}}^ n\), g: \({\mathbb{R}}^{n(N+1)}\to {\mathbb{R}}^ n\) are continuously differentiable. The author employs the contraction mapping principle to show that, under suitable hypotheses, if \((1_ 0)\) has a unique solution then \((1_{\lambda})\) has a unique solution for \(\lambda\) sufficiently small. For system (2), results concerning the existence of isolated solutions are obtained. These results extend known results for systems with linear boundary conditions.