Certificate Program

in

Computational Science

within the

Department of Mathematics & Statistics

at

Old Dominion University

Computational Science, and more broadly Computational Science & Engineering (CS&E), is an interdisciplinary research concentration that has been prospering under broad federal initiatives in High Performance Computing & Communication (HPCC) since about 1992, and earlier in isolated pockets of expertise. CS&E is most strongly established in the national laboratories, in federal agencies concerned with science and technology, and in the science and engineering professional societies. These organizations sponsor graduate fellowship programs, interdisciplinary conferences, and research funding initiatives, in addition to investing internally to build up their own CS&E capabilities. Publishers have begun to target CS&E as a niche with journals and book series. Increasingly, universities are catching up with federal and professional initiatives by offering graduate and undergraduate majors, minors, and certificates in CS&E, both within and between traditional academic units.

Old Dominion University has several faculty members working in CS&E and has successfully placed a number of students who have fashioned their own CS&E programs inside of traditional majors in federal agencies. Following expected formal approvals, the Mathematics & Statistics Department will offer a Certificate Program in Computational Science for M.S. and Ph.D. degree candidates, effective beginning with current matriculants.

Early graduates of the program have taken employment at Argonne and Lawrence Livermore National Laboratories. Internships at leading national laboratories are an important part of the ODU program. During each of the summers of 2000 and 2001, there were six ODU HPCC graduate students or alumni at DOE laboratories. There are also currently six ODU HPCC graduate students with daily access and computer privileges at the Institute for Computer Applications in Science and Engineering (ICASE), at the NASA Langley Research Center.

Computational Science & Engineering is an interdisciplinary area that draws upon the traditionally distinct academic departments of computer science, applied mathematics and statistics, and one or more of the sciences or engineering, broadly interpreted. It is more than any of its component fields, and traditionally does not fit within the subject areas, conference and journal scopes, or reward structures of computer science, of mathematics, or of any of the science and engineering disciplines alone. An example of a research endeavor that would be considered CS&E is the development of a large-scale computer simulation code to investigate a scientific question that goes beyond the use of canned (commercial or freely available) software but requires familiarity with aspects of the simulation, such as computer architecture or algorithmics, that would ordinarily be considered to lie underneath an opaque interface to the scientific user. Approached from the other direction, an applied mathematician might develop an algorithm that adapted in a special way to knowledge of the physics and/or knowledge of the computer architecture to achieve much higher scalability, efficiency, or performance than a generic "brute force" approach ever could, resulting in an entirely new level of fidelity in the simulation. The Computational Science & Engineering curriculum is sometimes defined by the intersection of subjects needed in any large-scale simulation, independent of scientific origin, such as parallel computing, data visualization, error analysis, etc. However, it is better thought of as the union of subjects required in a particular large-scale simulation, such as the physics, the algorithmics, the coding implementation, the hardware implementation, etc. It is not necessary to be proficient in every science to be called a computational scientist; however, it is necessary to be "vertically integrated" in some science, from model creation down to critical interpretation of the ultimate computational results, to earn this certification.

Graduate students seeking the Certificate as part of their M.S. or Ph.D. in the Mathematics & Statistics Department will fulfill the standard requirements of those degrees. In addition, they will fulfill requirements particular to the Certificate Program. There are five requirements: in curriculum, in degree committee population, in internship, in colloquium participation, and in the topic of independent work.

Curricular Requirement. In addition to satisfying the requirements for the desired advanced degree in Computational and Applied Mathematics, students will elect a minimum of 12.0 credit hours of their 500-level and 600-level graduate coursework (outside of one-time topics courses) in computer science and in science or engineering applications. They will also elect the 3.0 credit hour course Introduction to Computational Science & Engineering. This course, offered at least once per year, consists of regular faculty, guest, and student lecturers present case studies in computational science & engineering.

Degree Committee Requirement. For doctoral students, at least one member of the doctoral committee must be from outside the university and working in CS&E, in the judgment of the Computational Science Departmental Advisor.

Internship Requirement. Every student, whether seeking an M.S. or a Ph.D. degree, who wishes to qualify for the Certificate in Computational Science must spend a minimum of ten weeks in an internship, paid or unpaid, at an off-site research organization engaged in CS&E. Examples in the immediate vicinity include the NASA Langley Research Center, the Thomas Jefferson National Accelerator Facility, the Newport News Shipbuilding Company, some branches of the Department of Defense, some for-profit defense and government contractors, some for-profit software, hardware, and network vendors, etc. Eligibility will be determined by the Computational Science Departmental Advisor. An application for eligibility must be approved before the internship begins and a report on the internship must be filed at its conclusion. Ideally, for doctoral students, the internship should be in the specific area of the thesis, but this is not a requirement. Ideally, the internship should be with the external member of the student’s doctoral committee, but this is not a requirement.

Colloquium Requirement. During a course of two or more years of graduate study in residence at ODU, there are dozens of graduate/professional seminars presented in the colloquia of the Mathematics & Statistics Department, the Computer Science Department, and other departments in the Colleges of Science and Engineering that would be considered CS&E. To qualify for the Certificate, a graduate student must attend at least six such seminars, do a brief websearch on the speaker and his/her project, and write a brief summary of the state of the art in the area of the seminar.

Independent Work Topic Requirement. Students in Mathematics & Statistics generally perform some form of independent research in pursuit of their graduate degrees. It is understood that students seeking the Computational Science Certificate should do at least a substantial portion of their independent work, including the thrust of their doctoral thesis for doctoral students, in the area of CS&E. In the course of their dissertation work, students must of course also make a significant and original mathematical or statistical contribution, according to the usual standards.

Students are required to take at least 12.0 credit hours outside of the Department of Mathematics & Statistics, as follows: