Background

Computational science is about using computers to analyze, understand and solve complex scientific problems. It is distinct from computer science which studies computers and computations.

Computational science covers a broad spectrum. At one end, some of it falls easily into the category of ``theory'', as exemplified by the theoretical condensed matter physics or quantum chemistry research that has been ongoing at Memorial for many years. Towards the other end of its spectrum, the acquisition, structure and analysis of huge data sets can pose extremely challenging computational problems. The analysis and processing of seismic data in our Department of Earth Sciences fall into this category. Imaging and image processing, carried out in MUN's Department of Physics and Physical Oceanography, constitute a similar field of endeavour. Somewhere in the middle is the very important field of numerical simulations. These simulations are often thought of as being somewhat distinct from both theory and experiment, and are commonly referred to as ``computer experiments''. The general methodology consists of representing a physical system via a set of equations which are then solved numerically, and the solutions used to ``measure'' various physical characteristics of the system via some numerical sampling technique. One illustrative example of this type of work is the simulation of the dynamics of a finite system of molecules. With a set of intermolecular forces which are assumed to be known, the dynamics of the system are calculated, as a function of time, via the solution of Newton's laws. Such work can be extremely useful, providing both qualitative and quantitative insights into many phenomena that are too complex to be dealt with by analytical methods and too difficult, expensive or even dangerous to study via experiment.

The size and complexity of problems of interest to computational scientists generally require powerful computational resources. In the modern era, these resources can be husbanded in a variety of architectures, from massive supercomputers to distributed computer systems. The study of such architectures and their efficient application to different scientific problems is, itself, a part of computational science.

In some cases, computational scientists develop their own computer codes. However, just as experimental scientists often build upon and use equipment developed and produced by others, practicing computational scientists often use software developed by others. It can range from general purpose software libraries such as NAG or IMSL, through somewhat more specialized tools such as MATLAB or Mathematica, to the huge, powerful and highly specialized software programmes such as those used in quantum chemistry or the pharmaceutical industry. Learning how to use these tools in an efficient manner, how to refine and enhance them, and how to intelligently interpret the results, is an essential part of the training.

Not only are the problems and the computer architectures complex; in many cases, so are the solutions to the problems. Largely for this reason, computational science relies heavily on visualization to represent problems and solutions.

It should be apparent that computational science can be an integral part of many branches of science, complementing and supporting theoretical and experimental studies. In many areas of research, making realistic simulations in order to understand complex problems requires a team effort that combines theory, experimentation and computing. Such a multidisciplinary approach is fundamental to computational science. Training and practice in presenting results to non-specialist colleagues will also be an integral part of the programme.


Next Page
Up to Table of Contents
Previous Page