Scientific computation

Overview

The SciCom group carries out research on methods and software for supporting applications in such areas as medical imaging, option pricing and hedging, portfolio optimization, biomedical modelling, computational neuroscience, and physically based animation.

Group members develop and analyze algorithms for solution of large sparse matrices, large-scale optimization, least-squares problems, and numerical solution of partial differential equations. These algorithms are then tested in a variety of application domains.

Each application brings specific computational difficulties and challenges. One of the challenges in scientific computation is to understand these computational difficulties and meet them with effective methods embedded in efficient software. Current collaborations with researchers and practitioners in finance, graphics, and medicine have helped the SciCom researchers to understand the computational requirements of these applications.

Past collaborations include extensive involvement with hydrogeological modelling, medical imaging, computational fluid dynamics and semi-conductor device simulation.

SciCom has been active in object-oriented design and implementation for software that is reusable across a variety of applications, particularly for sparse matrix solvers. Software packages such as Sparspak and Watsit, developed by SciCom members, have been licensed by a variety of industrial and academic users.