Operations Research Center
The 'Center' of Attention
Founded as an ad-hoc program by Philip Morse, the Operations Research Center at MIT prepares to celebrate 50 years at the top of the class of OR grad school education
By Peter Horner
|
The Operations Research Center at the Massachusetts Institute of Technology is, however, MIT's "oldest, continuously operating interdepartmental, interdisciplinary degree-granting program." That's quite a mouthful — and quite a feat — considering the long and storied history of MIT. What makes the story of the ORC all the more remarkable is the fact that it began 50 years ago as an ad-hoc program in the office of a physics professor.
Of course, the physics professor in question — Philip McCord Morse — was no ordinary OR guy. Widely known as the "father" of operations research in the United States, Morse was one of founders of the Operations Research Society of America and served as ORSA's first president in 1952. He literally wrote the book on the subject, "Methods of Operations Research," co-authored with George Kimball. The text (recently reprinted by Dover Publications) has introduced countless students and practitioners to the field since it was first released to the public in 1951.
Morse was unquestionably an intellectual giant with remarkable insight — a man who was well ahead of his time — but it's doubtful even Morse could have foreseen that the ad-hoc program he hatched in his campus office would turn out to be the gold standard by which all other university OR departments are measured.
Time to Celebrate
On April 24-25, ORC alumni and current and former MIT faculty associated with the program will gather in Cambridge, Mass., to celebrate the 50th anniversary of the Operations Research Center. The two-day fest will begin on Saturday, April 24 with a private symposium for ORC alumni and MIT faculty featuring an all-star line-up of speakers. (See story on page 34). On Sunday, April 25, the celebration will continue with an afternoon of ORC-inspired sessions presented as part of INFORMS Practice Meeting in Cambridge. All meeting registrants are invited to the Sunday sessions. (See story on page 32).
"We're celebrating 50 very successful years of what is now the country's largest doctoral program in operations research," says Richard Larson, a former co-director of the ORC who heads up the anniversary festivities.
With about 40 Ph.D. students in residence, the ORC is certainly large in terms of dedicated OR departments. The impact that the ORC has made on the OR profession and INFORMS (and its predecessors, ORSA and TIMS) is immeasurable. MIT faculty, and students and alumni of the ORC, dominate the honor roll of INFORMS officers and award winners going all the way back to Morse and the founding of the Institute.
John D.C. Little, the first in a long line of Ph.D.s to come out of the ORC, succeeded Morse as director of the ORC in 1968, and later served as president of both ORSA (1979) and TIMS (1984-85) and was elected the first president of INFORMS following the merger of the two organizations in 1995. Tom Magnanti, a former ORC co-director and now the dean of MIT's School of Engineering, served as president of ORSA (1988) and INFORMS (1999). Larson, another former president of ORSA (1993), was voted president-elect of INFORMS just last year. The list goes on and on. Literally and figuratively speaking, there's a lot more where they came from.
When you take into consideration all of the MIT faculty, alumni and students who have been associated with the ORC over the past 50 years, and add up their contributions in terms of research, practice, teaching and service, it is impossible to overestimate MIT's collective influence on the OR profession and INFORMS.
In the Beginning
In his autobiography, "In the Beginnings: A Physicist's Life," Morse recounted how a growing interest in OR among faculty and students at MIT prompted the administration to appoint him as chairman of a Committee on Operations Research "to coordinate education and research in the subject." The year was 1952, the same year Morse was voted the first president of ORSA at the society's founding meeting held at the Arden House in New York.
"By 1955 there was enough interest in O/R at MIT to justify looking for money for the support of graduate students who wished to specialize in it," Morse wrote in his autobiography. "This time it was obvious where to go for funds; the armed services employed many O/R experts and were interested in having us train more."
Morse contacted Tom Kollian, then the civilian head of the Office of Naval Research, and landed two contracts — one from the Navy, the other from the Army, each worth a few hundred thousand dollars.
"Armed with our initial contracts," Morse wrote, "I was able to ask the Institute to authorize the formation of an Operations Research Center, with offices to house the student assistants and with an administrative staff to keep track of the student's work, to monitor expenditures, and to write the necessary reports to the contracting agencies."
Morse added: "Research in O/R was still viewed with skepticism by many faculty members, but since we now had money to support students and there were students who wanted to join the Center, there was no active opposition. Students from all over the Institute were attracted."
Horizontal vs. Vertical
From the beginning, the interdisciplinary, interdepartmental nature of the ORC set it apart from traditional academic departments. Instead of the vertical or "stovepipe" infrastructure associated with most departments, the ORC operates horizontally across different schools and departments. At first, the physics and mathematics departments were most likely to be involved; now it's the School of Engineering's Civil and Environmental Engineering Department and Electrical Engineering and Computer Science Department, as well as the Sloan School of Management. The interdisciplinary, interdepartmental approach is reflected in the ORC's now traditional leadership by co-directors — one representing the School of Engineering, the other the Sloan School.
In the beginning, the ORC could not admit graduate students directly into the program. Prospective students had to first gain admission into a regular department and then seek admission into the ORC through a petition process. The procedure required the student to find at least three faculty members from two or more academic departments who were willing to supervise the student's doctoral research.
As Morse noted in his autobiography, "The Center had no faculty of its own; each participating faculty member was a member of some department and had been appointed and would be promoted at the pleasure of that department, not necessarily in response to the desires of the O/R Center. This meant that I, as director of the Center, had to convince the various department heads and deans that it was advantageous to their departments to have a member with interest and experience in operations research. In most cases, I managed to do so. I still feel that the advantages of our interdepartmental nature outweighed that additional work."
As the program evolved and became larger and more successful, the MIT administration empowered the Center to admit master's and doctoral students directly into the program without requiring the Sloan School or the School of Engineering to sign off on the deal.
"Over a period of time, the OR Center very cleverly institutionalized something that had been an ad-hoc and rare procedure," says Larson, referring to the student petition process. "That's how the ORC became a regular degree-granting institution within the university."
Incentive and Reward Structure
About 35 faculty members voluntarily find their way to supervise ORC students' research, even though the university does not compensate them for their time. So what's the incentive and reward structure?
"It's very simple," says Larson, a professor of engineering systems and of civil environmental engineering at MIT, and a participant in the ORC program in various capacities since the mid-1960s. "The ORC students are unsurpassed at MIT in terms of their quality, personality, energy, perseverance and creativity. Faculty members are just drawn to them."
The ORC may not support any faculty or offer any courses, but ORC students face a very structured, rigorous curriculum that involves both theory and applications as well as an OR practicum. At MIT, any academic subject (courses are called subjects at MIT) has to be taught through a department. If a faculty member wants to teach an "OR" subject, they first and foremost have to convince their academic department that the subject they want to teach is useful to that department. "If it's also interesting and useful to the ORC program, that's considered a bonus," Larson adds.
Over the years, faculty members have enjoyed considerable success selling their departments on the value of subjects like queueing theory and linear programming. As new subjects gained favor, the ORC's erstwhile ad-hoc academic program evolved over many years into the rigorous, structured curriculum it is today. Whether they're theoreticians, applications people, optimizers or probabilistic folks, all ORC students take the same set of core academic subjects and undergo the same general exam structure.
"They all have to take this set of courses involving probability and optimization that gets them up to research knowledge in one of those, and almost up to research knowledge in the other," Larson says. "In addition, they have to take an operations research practicum, which is usually in the summer, where they actually have to do applied operations research. Even the theoreticians have to take the practicum and do applied work."
Once the students meet the core requirements, however, "they can blossom as they want," Larson says. "They can be an Ed-Kaplan-HIV-modeler. They can be a Dimitris Bertsimas-deep-generalist-person. They can be a Rob-Freund-interior-point-theoretician. They can be a Cynthia-Barnhart-yield-management-airline-applications-person."
And some become a Richard-Larson-urban-OR-person.
Best and Brightest
The ORC's interdepartmental, interdisciplinary approach — backed by MIT's world-class academic reputation and facilities — attracts some of the best students in the world. Once they are brought together under one roof (every grad student gets his or her own cubicle at the ORC), the students tend to bond together. Instead of competing with one another, Larson says they help each other with subjects and projects. They brainstorm ideas, they socialize together, they have bowling nights, they organize group skiing trips.
"At most of the academic departments I'm familiar with, individual faculty members operate by themselves with their own cadre of graduate students," Larson says. "They may serve on committees and that sort of thing, but that's about it. At the ORC, there's mentoring of untenured faculty. We reach out and draw them in. I don't know of any academic department that has the sense of camaraderie and the shared purposed throughout that the ORC center has.
"We get very talented, very energetic students, the best in the world," Larson continues. "They come here and do all of this amazing stuff. And then they leave and go on to even greater things. It's a family tradition that's passed along from generation to generation."
The list of ORC alumni who have gone forth and made a major impact on the field of operations research — people like Ralph Keeney in the area of decision analysis and Bruce Golden in the area of network optimization to name just two — is truly staggering.
The Spirit of Morse
Many of those alumni, as well as current and former faculty, will be gathering at the Cambridge campus once more to toast the ORC on its 50th birthday. Does the sprit of Phil Morse still pervade the place?
"It persists in individuals," says Larson, who has lobbied MIT to rename the ORC in honor of Morse. "It certainly persists in me and some of my colleagues like Amedeo Odoni and Arnie Barnett. OR was born of physicists as an empirical science, and it really is a branch of applied physics rather than applied mathematics or engineering.
"Look at Morse and what he did in OR. He was a physicist who was both an experimentalist and a theoretician. What do they do? They go out in the real world. They observe and collect real data. They come back to their office or laboratory and try to construct a mathematical model of what they've observed. They then test the model's predictability. They go out and test that model versus the real world again, and through an iterative process they create a pretty good mathematical depiction of the physical phenomena you're trying to understand. Operations research, from that point of view, is a branch of applied physics. We are trying to understand the physics of the world in which humans interact with technology and with each other.
"Operations research was born of applications. All the major theoretical work that has been done in OR — queueing theory, graph theory, linear programming, dynamic programming — all of these things were driven by the need to solve real problems. I don't know of any major OR breakthrough that was driven by a theoretical mathematician.
"If OR is only about applied math, then OR will die. Think about it. OR is research on operations. How can you do research on operations at the blackboard in your office? The challenging mathematical problems come from the wonderful world around us, which is infinitely more complex than what we can invent in our office."
|
"Whether it's HIV, bioterrorism or March Madness that Ed is modeling, it's all driven by real problems and real data," Larson says. "Morse's influence has not died. It exists in the practicum that every ORC student has to take. It exists in the culture at the OR Center that has produced people like Ed Kaplan and so many other outstanding graduates over the last 50 years."
Peter Horner is the editor of OR/MS Today.
OR/MS Today copyright © 2004 by the Institute for Operations Research and the Management Sciences. All rights reserved.
Lionheart Publishing, Inc.
506 Roswell Rd., Suite 220, Marietta, GA 30060 USA
Phone: 770-431-0867 | Fax: 770-432-6969
E-mail: lpi@lionhrtpub.com
URL: http://www.lionhrtpub.com
Web Site © Copyright 2004 by Lionheart Publishing, Inc. All rights reserved.