Parallel alternating explicit implicit domain decomposition algorithm

Abstract

Explicit-Implicit Domain Decomposition (EIDD) is a class of globally non-iterative, non-overlapping domain decomposition methods for the numerical solution of parabolic problems on parallel computers, which are highly efficient in terms of both computation and communication for each time-step. The goal of this thesis was to investigate a parallel implementation with the novel domain partition for Alternating EIDD (AEIDD), which is used to solve general parabolic problems. We achieved increased parallelism than was previously implemented, while maintaining the stability, efficiency and accuracy of the AEIDD algorithm.