On a numerical method for a homogeneous, nonlinear, nonlocal, elliptic boundary value problem (Q629285)

A numerical method is developed for a class of nonlinear, nonlocal boundary value problems of the form \[ \begin{cases} -\alpha (\int_0^1 u(t) dt) u'' (x) + (u(x))^{2 n + 1} = 0,\;x\in(0,1),\\ u(0)=a,\quad u(1)=b,\end{cases}\tag{\(*\)} \] where \(a\) and \(b\) are positive constants. For this purpose, the following nonlinear auxiliary boundary value problem is considered: \[ \begin{cases} -\alpha (q) u'' (x) + (u (x))^{2 n + 1} = 0, \;x\in(0,1)\\ u(0)=a,\quad u(1)=b.\end{cases}\tag{\(**\)} \] Some results concerning a priori estimates and existence and uniqueness of solutions of (\(**\)) are established by means of Schauder's fixed point theorem. These results are effectively used to obtain results concerning existence and uniqueness of solutions of (\(*\)) by Brouwer's fixed point theorem. The problem (\(*\)) is then discretized and the existence of a unique solution of this discrete nonlinear problem is established. It is shown that the numerical procedure converges with an error \(o(h)\). Some examples are considered to illustrate the numerical process.