scientific article; zbMATH DE number 1124499

From MaRDI portal
Publication:4378538





The authors consider the following discrete boundary value problem \[ x(k+1)=f_k(x(0),x(1),\cdots,x(k)),\quad k=0,1,\dots,N-1\quad h[x]=0\tag{1} \] where \(x(k)=(x^1(k),\dots,x^n(k))\), \(f_k:\mathbb{R}^{n(k+1)}\to\mathbb{R}^n\) and \(h:\mathbb{R}^n\to\mathbb{R}^n\). For (1) the existence and uniqueness of solutions are investigated. This includes provision of a priori sufficient conditions which guarantee the convergence of Picard's iterative scheme to its unique solution. The Picard's method is then applied to perturbed boundary value problems. Further, the authors prove the monotone convergence of Picard's scheme. Next, sufficient conditions which ensure the convergence of quasilinear methods are provided. Finally, the authors discuss the computational aspects of quasilinear methods on a floating point system. This includes necessary and sufficient conditions for the convergence of approximate quasilinear method, an oscillatory state, and the stopping criterion.











This page was built for publication:

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4378538)