University of Wisconsin - Eau Claire

Computational Science Minor

UWEC MULTIDISCIPLINARY PROGRAM

What is Computational Science?

Computational Science is an interdisciplinary program that integrates various aspects of traditional natural and physical sciences (such as biology, physics, geology, and chemistry), applied mathematics, and computing. Computational science is not the "science of computing"; that is, computational science and computer science are completely different subjects. Of course, certain aspects of computer science do provide an important foundation for the computational scientist. In particular, problem solving techniques, programming in some high level language (such as C, C++, Java, and Fortran), programming in some symbolic/numeric application language (such as Maple, Mathematica, and Matlab), and an understanding of high performance computer architectures are all fundamental principles from computer science that are also important for the computational scientist.

Computational science is also not just "traditional science" with some ideas worked out via a computer program. Such an approach usually leads to a nice scientific concept implemented with inferior numerical techniques on an architecture poorly suited to the problem (or simply poorly implemented). Such an approach often also misses the value in using symbolic and numeric applications to "prototype" the scientific ideas first in order to gain a deeper understanding of the computational issues, or even to visualize the results of those ideas.

Computational science is not the end result of three separate "courses of study": science, applied mathematics, and computing. Again, such an approach misses the need to integrate all three at all times. A true computational scientist does not first wear his "science hat". Then having worked out all the modeling issues, he puts on his "numerical recipes hat". Then having selected his favorite mathematical method, he puts on his "computational hat" and implements and runs the program. On the contrary, the true computational scientist wears all three hats at the same time, using knowledge and experience gained in each area to improve/reject ideas from each of the other areas in an iterative process that eventually (hopefully) leads to a numerically robust, computational efficient, scientifically meaningful solution.

And computational science is not an "exact" science. One should not expect closed form solutions with guarantees of optimality or exactness in typical computational science problems. A typical problem in computational science has no known exact answer, and maybe even no way to be sure that any answer obtained by some technique is correct. It is for precisely this reason that the computational scientist must understand the numerical issues of the mathematical solution selected, and must also continuously ensure that the results obtained using the scientific model are still meaningful.

To train an individual to be a computational scientist therefore requires a new truly interdisciplinary approach, one that does not stress any one component at the expense of the other two. A typical course, or course sequence, should blend an actual scientific problem with a study of some possible numerical solution techniques (without avoiding a consideration of convergence and error) and the associated computational implementation using either a traditional programming language or a symbolic/numeric application.

The Computational Science Minor at the University of Wisconsin - Eau Claire is an interdisciplinary program consisting of 24 credits and involving the active participation of the Departments of Biology, Chemistry, Computer Science, Geography, Geology, Mathematics, and Physics & Astronomy.