The AARMS Collaborative Research Group in Numerical Analysis and Scientific Computing

stjohnsmun Group Administrator: Ronald D. Haynes (Memorial University), contact rhaynes AT mun dot ca


“Domain Decomposition Methods for PDEs" Short Course + Collaborative Workshop, organized by the AARMS Collaborative Research Group in Numerical Analysis and Scientific Computing, Aug. 4, 2015 — Aug. 8, 2015 - Halifax, NS, Canada

The use of computational methods to treat mathematical models in science and engineering is widespread. Such models often involve PDEs, and the efficiency of these algorithms on modern high performance computing systems relies on the ability to parallelize the computations. The aim of this workshop is to provide an introduction to the state of the art in theory and practical applications of domain decomposition (DD) methods for PDEs. DD methods are a divide and conquer approach to solving PDEs, splitting the global problem into small pieces whose solutions are obtained on individual processors or cores. These local solutions are then recombined to give a solution to the global problem.

The program will begin with a two-day short course given by Prof. Martin Gander (Geneva). Prof. Gander is an internationally recognized leading expert in Schwarz methods - a class of DD methods for steady state and time dependent PDEs. Prof. Gander has taught similar courses in Europe and Asia. The middle component of the program will focus on presentations by researchers whose work may benefit from the use of DD methods for PDEs arising as mathematical models in practical applications. The final segment of the program will feature a workshop format in which breakout teams will investigate the process of introducing DD techniques into the numerical simulations that arise in the applications identified earlier. This segment will also include several talks by researchers working in DD methods.

The organizers invite participation by researchers in DD methods for PDEs, researchers working in applied PDEs whose work may benefit from the use of DD methods for PDEs, students and post docs. The program will have room for 5 - 8 talks by researchers in application domains. For researchers in DD methods for PDEs, the primary form of participation will be through the collaborative workshop teams but there will also be room for a limited number of contributed talks.

Confirmed DD experts include David Keyes (Kaust), Victorita Dolean (Nice) and Felix Kwok (Hong Kong Baptist).

Registration will be required. Some funding for graduate students will be available. Further details and registration information will be available shortly here. Email: rhaynes AT mun.ca for more information.