Advanced Computing in the Age of AI | Saturday, August 13, 2022

OSC’s Ralph Regula School of Computational Science: Five Years Later 

<img style="float: left;" src="" alt="" width="57" height="57" />The Ralph Regula School of Computational Science at the Ohio Supercomputer Center is a statewide virtual school focused on the exciting new area of computational science – the use of computer modeling and simulation to solve complex business, technical and academic research problems. The authors describe why the school was founded, its operation, and lessons learned over the past five years. The school directly address some of the most pressing problems limiting the use of high performance computing and the adoption of digital manufacturing technology.

Recognizing that a workforce well-versed in the emerging field of computational science could jumpstart the Buckeye State's anemic rust-belt manufacturing economy, the Ohio Supercomputer Center and Ohio Board of Regents opened the virtual doors of the Ralph Regula School of Computational Science in 2006.

Their aim was to help build a workforce pipeline in the STEM disciplines of science, technology, engineering and math that would be well versed in computational science. This, they believed, would boost scientific discovery, innovation and increased efficiency in engineering design and development.

The Ralph Regula School of Computational Science at the Ohio Supercomputer Center is a statewide virtual school focused on the exciting new area of computational science – the use of computer modeling and simulation to solve complex business, technical and academic research problems. The school was named after former Ohio Congressman Ralph Regula, who had a public service career that spanned more than four decades.

A respected policymaker, Congressman Regula (shown here) has long recognized the importance of science education and supported these initiatives in Ohio.

In this article, we take stock of where the Ralph Regula School is five years later. Some significant successes have been recorded, while others have taken more time to realize, and much remains to be done.

The need is identified

In early 2005, officials at the Ohio Board of Regents and the Ohio Supercomputer Center held deep discussions on how the higher education community could play an even more impactful role in helping the Buckeye state turn around its manufacturing economy. Multiple sources pointed out the direction they needed to take:

  • U.S. Bureau of Labor Statistics predictions for 2006 to 2016 showed computing-related occupations as the quickest growing among all "professional and related occupations."
  • A 2004 Council on Competitiveness survey reported that 97 percent of major companies could not function without high performance computing and computational science.
  • A 2005 report from the President's Information Technology Advisory Committee stated, "Universities and federal R&D agencies must make coordinated, fundamental, and structural changes that affirm the integral role of computational science."
  • A 2008 Federal Computer Week article highlighted Bill Gates' lamentations that the lack of emphasis on STEM disciplines in America were detrimental to U.S. global competitiveness. In response, OSC's Ohio officials concluded: "…supercomputing horsepower is meaningless without an educated workforce to run those systems. Sadly, U.S. education is not producing enough leaders in computational science, the interdisciplinary field that uses computers to simulate and model complex problems."

More recently, Jon Riley, Vice President, Digital Manufacturing at the National Center for Manufacturing Sciences, noted in a Digital Manufacturing Report blog "…there are currently 600,000 manufacturing job openings in our nation today that have not been filled because manufacturers have been unable to find the properly skilled applicants. We need to be training our current workforce and developing our future workforce."

For the Ralph Regula School to address this pressing situation, several major hurdles would have to be tackled:

  • Faculty expertise – We found that individually, most institutions were typically unable to develop comprehensive programs in computational science. There are three primary reasons for this: Their curriculum is structured (and divided) along disciplinary lines; they often lack critically important faculty expertise in one or more disciplines; and they are unlikely to have more than a relatively small number of students interested (at present time) in this field.
  • Cyberinfrastructure – Most institutions lacked access to the necessary cyberinfrastructure (hardware, software, connectivity and middleware tools/systems) that was essential for delivering computational science education curriculum and linking to cyberinfrastructure research.
  • Curriculum – There were no widely accepted national guidelines for undergraduate computational science education. This is because computational science is an inherently inter-disciplinary field and, therefore, few departments felt it was their role alone to create a curriculum.
  • Institutional Recognition – Issues related to setting tuition charges, sharing tuition income, sharing human and other resources, awarding transcript credits and awarding degrees often seem intractable when multiple institutions are involved, particularly when they include private and public institutions.

Launch of the Ralph Regula School

Our solution to these issues was the creation of a statewide, virtual school of computational science, named after Ohio Congressman and education advocate Ralph Regula. The virtual nature of this program allowed students at various colleges and universities to combine classes from their home campus with those from any other participating institution. The idea was that one campus might have strong faculty/programs in the biological sciences, another campus in chemistry, and yet another in physics. Each would offer classes at a distance in their strong disciplines and encourage their students to sign up for courses at other campuses in less strong or nonexistent disciplines. At the same time, pooling students from several campuses, at least in the beginning, would fill most sections of the classes offered.

Creating such a shared curriculum is difficult enough when it involves multiple departments in the same institution. The problems are magnified when multiple institutions are involved and are complicated by differences in academic schedules, tuition charges, academic mission, and student populations. Yet initially there was a cadre of dedicated faculty across at least nine institutions that was willing to expend the effort to create a multi-institutional program.

With financial support from the Ohio Board of Regents and a National Science Foundation grant, we were able to work with those faculty members to reach a consensus on a set of core competencies in computational science and a draft model for inter-institutional collaboration to implement the program. The model was then taken to the academic leaders at each of the participating institutions for their consideration. After extensive discussions and revisions to the initial draft, the provosts at all nine institutions signed the inter-institutional agreement and each of the institutions worked their way through the lengthy academic program approval process. The principle idea that carried the program forward was the unique opportunity it created for the students while allowing the institutions to share the costs of program implementation.

Another of the keys to a successful launch of the multi-campus project is the ability to maintain communication with and access to cyber resources. Ohio was uniquely suited for such a challenge, as the Ohio Academic Resources Network (OARnet) had just lit one of the most extensive statewide dark fiber networks in the nation in 2004. The network backbone covered 1,850 miles and linked 90 of Ohio's institutions of higher education with a high-speed broadband network that allowed for very high quality video services, data transfers and linkages to OSC's computational resources. And the Ohio Learning Network would be able to assist with the delivery of distance education programs for those courses that were not offered by the home institution.

The school relies on participating colleges and universities to confer degrees and certificates and offer their expertise toward the following goals:

  • To implement a multi-institutional, interdisciplinary undergraduate minor in computational science
  • To cultivate and maintain curricula standards for computational science degree programs and certificates
  • To provide assistance (with OLN's help) in using technology to deliver courses and programs in the most convenient and effective way for students
  • To create standardized certificate programs to create workforce knowledge and skills valued by industry
  • To coordinate with industry ensuring that insights gained in the workplace enter the curriculum as quickly as possible
  • To support innovative ideas for strengthening program effectiveness, such as a Computational Co-Op program that would make it easier for students to work directly with business and industry while actively pursuing a degree.

Baccalaureate minor program

In 2007, the Ralph Regula School and nine charter colleges and universities rolled out a baccalaureate minor program developed through National Science Foundation (NSF) funding. Because of the inter-institutional nature of the program, the minor program was designed to be competency-based. The instructional materials associated with each competency or subset of competencies was then embedded either in existing courses on each campus, in new courses at one or more campuses or as stand-alone modules that might be taken at distance and at the time that each student is ready for them.

Competency would be demonstrated through one or more assessments of the student's abilities on a combination of exams and projects. The competency-based approach was preferred because it gives curriculum flexibility to participating programs and gives employers some assurance of the working knowledge of each graduate.

Further, a business advisory committee reviewed the draft competencies created by participating faculty members. The year before, committee members had confirmed the majority of the recommended competencies and offered some advice on topic emphasis and breadth, providing the basis on which faculty developed instructional modules that would meet the competencies.

The computational science minor enables science and engineering majors to apply computational tools to the problems in their discipline.

As such, the minor was broken into three broad categories: Prerequisites; the Computational Science Core Competencies; and the Discipline Specific Competencies. The goal of the Core competencies was to establish a foundation that can be leveraged in the Discipline Specific Competencies that are directly applicable to the practice of the student's chosen profession.

A committee reviewed the various options for the sharing of materials, instruction, and revenues associated with cross-registered students. Recommendations in this regard were made to the participating institutions on institutional procedures.

Associate degree program

To further bolster Ohio's educational pipeline, we worked with several Ohio community colleges on another NSF project to create an associate degree concentration in computational science for community college students majoring in biology, chemistry or physics. We developed courses and materials, professional development for high school and community college faculty, and a model for a shared program that could be replicated nationally. The grant also allowed for the training of faculty teaching at other Ohio community colleges and increased our group of collaborative universities around the state.

Stackable certificate program

Finally, Ralph Regula School officials worked with educators and manufacturers to develop a stackable certificates program, which provides current workers the opportunity to expand their skills in computational science. With this program, we reached out to a very diverse workforce, in terms of background and educational attainment: traditional undergraduate students; currently employed individuals who have completed college or have some college but have little or no experience in computational science; and experienced engineers and scientists with college degrees that may understand the fundamentals of modeling and simulation, but have not been able to apply those skills in a way that lends competitive advantage to their employers.

The first two certificates we designed introduce students to simulation and modeling, the mathematical and algorithmic methods used to model systems, programming concepts and scientific visualization. Students also take one discipline specific course, depending upon their major or interest. There are currently courses in computational biology, chemistry, and physics.

The more advanced certificates apply techniques associated with our focus on polymer science. Based on our interactions with PolymerOhio and its members, we have formulated three advanced certificates in computational science related to modeling and optimization of polymer manufacturing processes, molecular modeling of polymeric materials, and plastics part design. The main goal of these courses will be to train industrial researchers in the use of the latest modeling tools to analyze, design and optimize polymer processes, understand the effect of processing on part properties, and how to best select the desired material for a given set of mechanical performance requirements.

Earlier this year, the workforce training arm of Sinclair Community College in Dayton partnered with the workforce arm of Columbus State Community College and the Ralph Regula School to begin offering a pilot training program geared toward workers in fields such as financial services, insurance, technology, engineering, and industrial science. The program, known as Comp-U-Science, allows employees to take free online courses in enhanced modeling, programming, algorithm and computational skills. Funding for three courses is being made available at no cost to employers by an NSF grant. The coursework consists of introductory courses in modeling and simulation, computational methods and programming, and algorithms. Those who complete the program will receive a certificate in Comp-U-Science from Sinclair and Columbus State.

"I am a professional engineer and such tools are helpful in making me more productive," noted David E. Guza, a student in the certificate program and a professional engineer with Applied Engineering Solutions in Dublin, Ohio. "[The program] provides a form of accreditation that supports and enhances personal professional credibility with customers and colleagues."

Guza indicated that he was looking forward to more courses, specifically interested in certificate training in the areas of advanced manufacturing, new methods of rapid prototyping and new product development.

Tie-ins with Blue Collar Computing, MEP, NDEMC, etc.

The Ralph Regula School was built, in part, upon a foundation laid by several other OSC innovations and strengthened by its participation in other nationwide efforts.

With support from the Ohio Board of Regents, OSC launched the Blue Collar Computing (BCC) program in 2004 to provide companies with innovative, advanced computational technologies that would allow for the virtual development of new and improved products, such as cars, pharmaceuticals and financial products. BCC provides computational resources, hardware, training, software and expertise to industrial clients to enhance their competitiveness.

In 2010, OSC and PolymerOhio launched the Polymer Portal, an initiative designed to enhance productivity for small- and mid-sized companies (SMEs) by providing affordable access to advanced modeling and simulation capabilities. The program is funded by the National Institute of Standards and Technology's Hollings Manufacturing Extension Partnership (NIST MEP). User access to the program is available at the website and supported through software training and the guidance of PolymerOhio's qualified industry experts. By focusing on a specific sector within a specific region – the polymer industry in Ohio – this project is taking an approach designed to produce tangible, meaningful, and manageable results. NIST MEP is interested in the potential to apply learning from this project on a broader basis across other manufacturing industry sectors as facilitated by the nationwide MEP network.

In 2011, the White House awarded a five-year, $5 million grant to The Council on Competitiveness and several partner organizations, including OSC, to support the advanced manufacturing efforts of Midwestern small- and medium-sized manufacturers (SMEs). Along with significant support and participation from numerous Original Equipment Manufacturers (OEMs), the Economic Development Administration public-private venture formed the National Digital Engineering and Manufacturing Consortium (NDEMC). The effort will introduce modeling, simulation, and analysis (MS&A) to Ohio, Illinois, Indiana and Michigan supply-chain manufacturers who do not alone have the resources to implement this technology into their workflow. OSC will serve as one of four resource providers, with a number of subject matter experts working with SME enterprises to help them adopt MS&A in their business process. OSC experts will extend specialized web portals and create additional portals derived from OEMs to assist manufacturing suppliers.

Lessons learned

There are several lessons that we have learned over the past five years. We must make a better case to industry as to the value of training a workforce skilled in computational science, especially in light of the challenging economic climate of recent years. We must take advantage of synergies available through partnerships with related technology initiatives, both local and national. Lastly, we must find ways to market the degree and certificate programs more effectively to potential students.

Better demonstration of ROI to industry – In difficult economic times, we need to present to businesses a strong case for investing in computational science education and training. They need to see very specifically how each individual organization is going to benefit from the outlay of assets.

For our part, we need to design and deploy additional advanced skillsets that fit within the time and staffing constraints faced by industrial clients. Up to now, our offerings have focused on a mix of theoretical and practical approaches typical of academic programs. Business leaders need to see that we can offer skills that relate to direct solutions for their specific industry and less on the way these solutions were derived.

Secondly, our courses need to represent an ever-closer alignment with available jobs. Businesses need workers that can provide their companies with immediate benefits arising from the training they receive.

Value-add partnerships – The resources that thus far have been invested in developing, promoting and deploying computational science education program have been put to good use, but are very limited. In order to continue growing our ability to develop and deliver such programs, we must look for additional support, but also find additional ways to partner with other initiatives that would lend support and funding to projects of mutual interest.

In one example announced in July, a national partnership of 17 institutions announced the Extreme Science and Engineering Discovery Environment (XSEDE). XSEDE, grown out of the former TeraGrid initiative, will represent the most advanced, powerful, and robust collection of integrated advanced digital resources and services in the world. NSF will fund the XSEDE project with $121 million over five years. In addition to many other goals of the project, we hope to work with universities around the country to formalize computational science education programs and prepare a future workforce that can apply computational modeling to solve challenging science and engineering problems.

We must also further develop opportunities to offer more computational science training options through the many web portal sites being developed to provide supercomputing access to industry. While the access to HPC resources is vital to helping American industry compete on the global market, access without the expertise in how to apply those powerful resources misses the mark.

More effective marketing to students ­– Marketing the programs to potential students has been a major challenge for the Ralph Regula School. After five years, only a few dozen students are actively engaged in the baccalaureate minor and associate concentration programs. Part of the problem lies in the diversity of the institutions involved using different terms to describe the programs or exhibiting major differences in media channels with which students can be reached on each campus. The answer seems to lie in engaging professors and enrollment management officials at each of the universities to actively promote the programs and the professional opportunities they offer. The professors have direct contact with the students in their disciplines who would most benefit from adding a computational science minor to their specialized area of study, and the enrollment officials have the best handle on message timing and the communication channels that best reach students on each campus.

An additional challenge lies in the interdisciplinary nature of computational science. We have found that the professors are our most effective resource for promoting the program to students. We must grow champions of computational science in each discipline that contributes to the field in order to reach the students who come to computational science from the various perspectives.

We also have used the employability factor to pique student interest. One approach to promoting the need for computational science has been to point out the industries, most with Ohio ties, which employ people with computational science skills and examples of how they use computational science. These include industry leaders, such as:

  • Procter & Gamble in Cincinnati, which used computational science to redesign their Pringles potato crisps so they wouldn't flutter off moving conveyor belts during production
  • Ford Motor Company in Cleveland and at several other Ohio sites, which used modeling to predict noise, vibration and harshness (NVH) performance of their powertrain assemblies
  • Medical Device Solutions in Cleveland, which used computational science to perform complex computational modeling to create new medical devices
  • GE Aviation in Cincinnati and other locations, which uses modeling of complex turbomachinery to create better, faster, more fuel-efficient jet engines
  • Military researchers at federal labs in Ohio, who used computational science to design energy-absorbing seats for armored vehicles to protect soldiers' ability to survive a mine blast
  • Nationwide Mutual Insurance Co. in Columbus, which uses risk modeling techniques by creating regional climate forecasts to estimate future losses
  • J.P. Morgan Chase & Co. in Columbus, which uses grid computing to develop mathematical models for pricing, hedging and risk measurement of derivative securities
  • Timken Corporation in Canton, which used modeling to improve NASCAR suspension system components and steering system geometries
  • Rolls-Royce Energy Systems, Inc. in Mount Vernon, which modeled the effects of jet engines striking birds in flight and engine behavior following the collisions

And, while the traditional education route has seen less-than-impressive enrollments so far, the initial offering of the workforce certificate program was wait-listed this fall. Officials at Sinclair Community College in recent months had to cut off enrollment at 33 students to maintain an acceptable student-teacher ratio in the Comp-U-Science certificate program.


The one unequivocal fact is that the need for computational science education is enormous and the potential economic return on investment is great. Industry needs more employees with acute computational science skills, and the American workforce, particularly in the Midwest, needs more employment opportunities.

Over the last five years, we've encountered several obstacles that stand in the way of meeting those needs – some we anticipated and several we had not. However, we've found that innovative approaches are available to help guide us around those obstacles. The biggest challenge that now remains before us is to scale up efforts in Ohio and across the nation to bring computational science to the tipping point where supply begins to meet the demand.

About the Authors

Dr. Steven I. Gordon is the Interim Co-Executive Director, Ohio Supercomputer Center and Director, Ralph Regula School of Computational Science.

Jamie Abel is the Media and Communications Director, Ohio Technology Consortium.

Add a Comment