August 1999
 |
||
 |
|
Does OR have a special territory, or has it died in the sense that every OR area overlaps, and is subordinate to, the rightful territory of another discipline? To paraphrase Mark Twain, the report of OR's death was an exaggeration. In fact, our track record of successes grows daily, and we see a bright future ahead. I believe OR may well become a leading field of applied science by the mid-21st century. Meanwhile, OR's major applications accumulate rapidly. For example, Industry Week ran an article in May about the rise of revenue management (or yield management) — worth billions to the airlines — now moving into hotels, rental car firms, entertainment and manufacturing. The 1999 Edelman competition, just completed in Cincinnati, further illustrates our growing impact. The winner, IBM, gained an estimated $750 million so far, with more expected later, from OR-based supply chain management. Another Edelman finalist, the Ford Motor Company's Visteon parts group, did creative manufacturing simulation that broke open a production bottleneck to increase output. Visteon only revealed how many extra hot-selling 4X4 sport utility vehicles and trucks were consequently sold over 18 months. But using unit-profit estimates from newspaper articles, anyone can calculate that the additional units probably were worth around $2 billion on the Ford bottom line. |
These examples are mere samples from the OR cornucopia. Many great projects don't show up in the Edelman competition or in publications. We know that classified OR analysis to guide military operations is widespread in the armed forces, that work considered proprietary is widespread in business, and that most practitioners are under pressure to turn out projects, not publicize them.
Its increasing strength and glowing outlook notwithstanding, OR would benefit enormously in the immediate future from greater visibility and appreciation. The steps to achieve this, roughly speaking, are I think:
- Clarify what OR is and does (map the OR territory).
- Settle in on just one name.
- Pass the word about OR to those who authorize, or could authorize, OR projects.
Why Map the Territory?
I see several good reasons to clarify what is OR. One reason is that we can sell OR more successfully when we can state what it is. We need to "brand" OR to sell it better. Another reason is that many outside of the OR community, who are only partly qualified or hardly qualified at all, would love to do our projects. In those cases we can't readily say that's work for the OR professional if we can't readily state what constitutes OR.
Also, everyone in OR can do better work when we combine forces to share expertise in methods and applications. The sharing takes place more freely when we recognize that we are in the same overarching discipline. Yet another reason is that clarifying for clients what we do helps demystify our field. The clients I have known wanted to understand the help they received with their important work-scene initiatives. They did not like ambiguity, mystery and black boxes.
The classical definition of OR is clear enough. After 50 years, however, the classical definition has become seriously out of date. We should now answer the question, "What is modern OR?"
Morse and Kimball
The evolution over the years of what OR means, starting with the famous Morse-and-Kimball (M&K) definition, is fascinating. I won't try to trace it here. There will be a panel on the topic, chaired by Carl Harris, at the upcoming Philadelphia meeting of INFORMS.
I do want to begin with the original M&K definition because this important piece of our history sets the stage perfectly. It comes, of course, from "Methods of Operations Research" (1951):
Operations research is a scientific method of providing executive departments with a quantitative basis for decisions regarding operations under their control.Before I propose a comparable definition of modern OR, let me detour to allow for further introduction, and especially to comment on steps two and three.
INFORMS' Broad Umbrella
Please note that INFORMS is home not only for OR but also for other fields and disciplines directly or indirectly affiliated with the management sciences. To set forth a definition of OR, therefore, does not in any way imply discouraging some individuals or groups from coming under the broad INFORMS umbrella. I know the board and staff join me in assuring you that all whose disciplines are among the management sciences are warmly welcome in INFORMS.
If names were wealth, OR would be enjoying an embarrassment of riches. You know the lineup: management science, operations analysis, operations evaluation, system analysis, decision technology, decision support, management technology, management analytics and so on. The niche names add to the confusion — recently things like marketing engineering and financial engineering. The catchiest niche name I suspect is Wall Street's rocket science, which the press loves. Perhaps the most exotic niche name is complexity science.
I don't object to those in a particular sector wanting their own banner. Our challenging collective problem, however, is that we lack recognition of our general name. We are like physicians with recognized names for their specialties but no recognized name for the overall field of medicine.
The INFORMS Public Information Committee, working with the INFORMS director of public relations, decided they will concentrate on the name operations research in outreach about OR to the press and public. That name, while not self-explanatory, still is by far the best established. Some day, perhaps, another name, one that requires less explanation, will take its place. Meanwhile we all can help the cause by mentioning wherever possible the same name — operations research — when we mean the whole OR profession or instances of it in niches. Thus, for example, mention that marketing engineering is a branch of operations research.
I should add that the abbreviation OR/MS is best interpreted to stand for operations research and the management sciences — the expansive, welcoming umbrella of INFORMS. It is not helpful, in light of our quest for greater recognition, to use it to mean operations research/management science, where management science is a synonym for operations research and thus is superfluous. Imagine how baffling it is to an outsider to discover that management science can mean either operations research or one of the management sciences or all of the management sciences.
In this soundbite era, if we hope to compete for visibility and brand recognition, we aren't likely to succeed by being cornflakes today and raisin bran flakes tomorrow. And we certainly shouldn't expect to succeed by being guess-what-I'm-called-today flakes.
Let the Word Go Out
The public relations program of INFORMS is making noteworthy progress connecting with the media. All of us are benefiting. The article in Industry Week about revenue management, for example, was full of quotes from experts referred to the reporter through INFORMS. The PR program publicizes many articles that appear in INFORMS publications, along with other newsworthy things like the Edelman competition. The PR committee does ask everyone involved in PR initiatives to please mention operations research explicitly when interacting with reporters.
If you have an opportunity outside of the PR program to represent the profession to the press or to our other constituents, current and prospective, you can help build brand recognition for OR by using the words operations research whenever practical.
What is Modern OR?
Once we stand together on calling our field OR, and getting the word about OR out to the world, we're ready to roll as soon as we agree upon what defines the field — as soon as we map OR territory.
Those old enough to remember early television no doubt fondly recall Professor Irwin Corey, "the world's foremost authority." His routine, invariably hilarious, consisted of taking a subject — the more mundane, the better, say marmalade — and giving a college lecture on it. The lecture, mostly double talk, always began like this: "What is marmalade?" That would be the only part of the lecture you could understand. After a round of deep lecturing, he would pause and say, "What was the question?"
I pose the question "What is modern OR?" wishing that professor Irwin Corey were here to shape the definitive answer. I only hope that my answer will be half as satisfying as his would have been.
Home on the OR Range
The classical definition staked out ground for the OR homestead, barn and corral. A modern definition would map the ranges of the OR ranch. Few of us frequently ride from one end of the ranch to the other. But all of us should know, and be prepared to explain, where the ranges go.
The classical definition covers, essentially, analysis that improves managerial decisions. We should add, first of all, analysis to improve other managerial functions — problem-solving, policy-setting, planning and forecasting, for example. Next we should add work in collaboration with non-managerial clients, such as engineers or consumers, to enhance decision-making and other functions comparable to managerial functions. Then we should add software development, underscoring the development of strategic decision-support systems. And we surely should add the development of new services and products related to the foregoing OR roles.
Those are the basic ingredients. To express them in a formal definition comparable to that of Morse and Kimball, I have in mind two guidelines. First of all, in the spirit of Occam's razor (prefer simplicity), I try to add only what is manifestly necessary. So I will focus primarily on what we do. Most of the things one could say about how we do it are, in my view, not defining. Thus, even though desiderata like having a holistic, system perspective, or unearthing the real problem, or living in the client's environment to understand the issues, or checking onsite that the data are correct, are signature for us, I would not incorporate them in a tight definition.
My second guideline is that I try to frame a statement complete and understandable for clients, reporters and others outside our field, as well as for newcomers to the profession. This causes me to extend substantially beyond a single short paragraph.
Rather than give up on offering a concise definition similar to M&K's, I divide the definition into two parts (see below). The first part, brief but self-contained, is a modern counterpart of the M&K definition. M&K wrote just one sentence. I wrote three sentences, the first paralleling the M&K sentence, the last two furnishing important amplification. The second part is a clarifying addendum.
Don't Forget Dodge City
The ranch couldn't survive without the ranch hands, food, supplies, churches, rail line and other support from Dodge City. Similarly, OR depends crucially on members of the OR community who create and extend the tools of OR, assemble the scientific knowledge of how organizations and decision-makers actually behave, train the next generation of OR professionals and educate future executives and managers who will be clients for OR services. I defined OR as Morse and Kimball did — as an applied-science field, with the definition focused on the applications.
Morse and Kimball's definition was paragraph one in a book of 158 pages. While I hope the proposed modern definition maps the territory accurately, it necessarily leaves much to be said. I will confine myself in the remainder of this article to a few questions that strike me as typical of what customers for OR services might ask. The answers are mine.
Q: How does OR differ from standard management consulting?
OR professionals are expected, individually or in teams, to know and select appropriately from all the applicable scientific methods and knowledge. OR is the highest-tech management consulting on the planet. It's the only field that trains its members to diagnose the problem, know the full array of treatments that work, and apply the most cost effective. Others outside OR either know few treatments or know lots of treatments but for only a few maladies.
Q: How does development of a decision-support system by OR differ from development by IT?
Unlike IT professionals who focus on hardware and software, OR professionals focus on the analytical heart of the system. Central to OR is deciding what input data is needed, how to convert input data into really useful decision-guiding information, and how to present that useful information so as to be acted upon most effectively. OR professionals, furthermore, are especially adept at helping the client move from a vague idea of a needed system to a detailed blueprint that draws on the most appropriate technology. They also give a client the benefit of keen interest in understanding and helping to solve the client's real problems. At the same time, OR professionals bring excellent IT skills.
Q: What are the most common misunderstandings about OR?
OR is not research (that is, not pure research — scientific work to build the world's store of knowledge) and is not limited to operations (usually understood to mean some subset of an organization's activities). The applications of OR are not limited to a particular type of organization, to a particular managerial function, to a particular organizational level, or to only managerial clients. The methods of OR are not limited to a small list — OR draws from the complete, now extensive, list of available scientific methods that are useful in OR applications. (Thus, for example, while quantitative modeling is a signature method, OR is not limited to quantitative modeling.) OR is not limited to work called OR by the professionals doing the work. OR is OR, whether or not called OR by the professionals on the scene, and, I believe, whether or not they do a decent job. If the job isn't done well, it's still OR — just bad OR. We aspire to stamp out bad OR.
Q: How does OR differ from such apparently similar specialties as industrial engineering and applied statistics?
When work that fits the definition takes place, that work is OR. Many times another discipline overlaps OR by including some OR in its scope. For instance, industrial engineering, an application field, includes some manufacturing OR. Applied statistics, a methods field, includes some OR applications where the tools of statistics are the most appropriate tools. These overlaps are normal in the world of technology. The key to understanding them is to recognize that it's not a matter of one or the other — you can and do sometimes find multiple disciplines at work simultaneously. Only the discipline of OR, however, has the common theme of drawing from the best of all technology to help improve decision-making and organizational performance universally. The insights that come from sweeping over the full spectrum of tools, organizations, managerial functions and organizational levels, create a powerful synergy that no overlapping niche discipline comes close to matching.
Welcome to OR territory!
Here for your consideration is a proposed definition of modern operations research, to update the M&K definition. The clarification is a proposed addendum. Definition Operations research (OR) is the application of scientific methods to improve the effectiveness of operations, decisions and management. By means such as analyzing data, creating mathematical models and proposing innovative approaches, OR professionals develop scientifically based information that gives insight and guides decision-making. They also develop related software, systems, services and products. Clarification OR professionals collaborate with clients to design and improve operations, make better decisions, solve problems and advance other managerial functions including policy formulation, planning, forecasting and performance measurement. Clients may be executive, managerial or non-managerial. These professionals develop information to provide valuable insight and guidance. They apply the most appropriate scientific techniques — selected from mathematics, any of the sciences including the social and management sciences, and any branch of engineering. Their work normally entails collecting and analyzing data, creating and testing mathematical models, proposing approaches not previously considered, interpreting information, making recommendations, and helping implement the initiatives that result. Moreover, they develop and help implement software, systems, services and products related to their methods and applications. The systems include strategic decision-support systems, which play a vital role in many organizations. |
References
- Morse, P. M. and G. E Kimball, 1951, "Methods of Operations Research," The Technology Press of MIT, Wiley, and Chapman & Hall, Cambridge, New York, and London.
Acknowledgments
I received generous help with this article from colleagues, especially members of the INFORMS Public Information Committee and INFORMS Task Force on Continuing Professional Education. I thank particularly Newt Garber, Dick Larson, Gary Lilien, Lew Pringle, Saul Gass, Sid Hess, Dennis Fuller, Carl Harris and Barry List.
Randy Robinson is working on a book for management about OR. A career OR practitioner, he chaired the ORSA-TIMS board that organized the creation of INFORMS and served as founding INFORMS executive director. He can be reached via e-mail at randyrobi@cs.com.
OR/MS Today copyright © 1999 by the Institute for Operations Research and the Management Sciences. All rights reserved.
Lionheart Publishing, Inc.
506 Roswell Street, 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 1999 by Lionheart Publishing, Inc. All rights reserved.