Computational method for a singular perturbation problem via domain decomposition and its parallel implementation (Q2366629)

Iterative algorithms for domain decomposition that reduce the boundary value problem \(\varepsilon u''=f(x,u)\), \(x\in (0,1)\), \(u(0)=u_ 0\), \(u(1)=u_ 1\), \(f_ u\geq \beta^ 2_ 0\), \(\beta_ 0=\text{const.}>0\), \((x,u)\in (0,1)\times \mathbb{R}\) to sequences of boundary value problems on subdomains are presented. The first algorithm is the Schwarz alternating procedure and the second one is highly suitable for parallel computing. The authors establish convergence properties of these algorithms and analyze a version of the second algorithm with two interfacial problems. Numerical examples are provided.