

![]() |
Convergence analysis of substructuring Waveform Relaxation methods for space-time problems and their application to Optimal Control Problems |
|
Author | ||
Director | ||
Defense | Thèse de doctorat : Univ. Genève, 2014 - Sc. 4745 - 2014/12/15 | |
Abstract | This thesis contributes to develop a new class of methods for the numerical solution of partial differential equations (PDEs) using space-time domain decomposition algorithms to ensure the use of different time steps in different subdomains. We first introduce and analyze new types of Waveform Relaxation methods based on the Dirichlet-Neumann and Neumann-Neumann methods, for parabolic and hyperbolic problems. The algorithms, formally termed as Dirichlet-Neumann Waveform Relaxation (DNWR) and Neumann-Neumann Waveform Relaxation (NNWR), generalize the use of substructuring methods to the case of evolution problems in a natural way. We finally propose an application of these methods for PDE-constrained Optimal Control Problems, solving the underlying forward and adjoint PDEs using a domain decomposition method. We apply and analyze the Dirichlet-Neumann and Neumann-Neumann methods on control problems and give the optimal choice of relaxation parameters for both the forward and adjoint problems in the steady as well as time-dependent case. | |
Keywords | Space-time domain decomposition — Waveform Relaxation — Dirichlet-Neumann Waveform Relaxation — Neumann-Neumann Waveform Relaxation — Optimized Schwarz Waveform Relaxation — Optimal Control Problems | |
Identifiers | URN: urn:nbn:ch:unige-461468 | |
Full text | ||
Structures | ||
Research group | Analyse numérique | |
Projects | Swiss National Science Foundation: 200020_149107 Autre: DIP | |
Citation (ISO format) | MANDAL, Bankim. Convergence analysis of substructuring Waveform Relaxation methods for space-time problems and their application to Optimal Control Problems. Université de Genève. Thèse, 2014. doi: 10.13097/archive-ouverte/unige:46146 https://archive-ouverte.unige.ch/unige:46146 |