FORCES, TRENDS, AND OPPORTUNITIES IN
MS/OR
ARTHUR M. GEOFFRION
University of California, Los Angeles, California
Operations Research, Vol. 40, No. 3, May-June 1992
This copyrighted material is not to be used without express
permission of INFORMS. Copying in whole or in part for use in the classroom
only is permitted, provided a complete copyright acknowledgment is given.
Copies of this article may not be used in any publication that will be
sold.
-
The purposes of this paper--a revised and extended version of the Omega
Rho Lecture given at the November 1991 ORSA/TIMS Joint National Meeting--are
to assess some important aspects of the current MS/OR situation and to
draw some conclusions about desirable future emphases. To these ends, it
identifies and discusses four forces of historic importance (the
microcomputer and communications revolutions, the dispersion of MS/OR in
industry, and academia's unbalanced reward structure), three major trends
(rapidly disseminating MS/OR tools, declining enrollments of native-born
students, and persisting management apathy toward MS/OR), and five outstanding
opportunities (ride the computer/communications revolution, support
dispersed practitioners, focus on the service sector, stress embedded applications,
and go into the quality business). An underlying theme is that the field
will flourish in proportion to how astutely individuals, organizations,
professional societies, and universities adapt to the changing realities
within which MS/OR lives.
The management science/operations research (MS/OR)
profession is where it is today largely because it has been subjected over
the years to certain powerful forces. The purposes of this paper are to
assess the most prominent forces and the trends which they have helped
to induce, and to derive some conclusions concerning how these forces and
trends can be harnessed for the long term good of the profession.
The first section discusses four of the most important forces impacting
MS/OR: the microcomputer and communications revolutions, the dispersion
of MS/OR in industry, and an academic reward structure strongly biased
toward theory. The second section considers three of the most significant
trends: rapidly disseminating MS/OR tools; declining interest of native-born
students in MS/OR studies, and persisting management apathy toward MS/OR.
These two sections include a summary of the consequences of each force
and trend for MS/OR as a profession and for individual professionals. These
consequences suggest directions in which individuals, professional societies,
universities, and other organizations may wish to move.
The third section, the principal focus of the paper, takes up five promising
opportunities for MS/OR to flourish in the kind of world implied by the
forces and trends described previously: riding the microcomputer and communications
revolutions; supporting dispersed practitioners; focusing on the service
sector; stressing embedded applications; and going into the quality business.
Most of the consequences for individuals wishing to capture the opportunities
envisaged here are imperatives: learn about . . . think about . . .
allocate some time to . . ., and so on. These accumulated imperatives
may seem overwhelming, especially for someone at an early career stage.
But there may be some comfort in recognizing that many of the suggested
activities count doubly or even triply in terms of career objectives, can
be dispatched by course work or occasional seminars, can be shared with
others if one works in teams, and can be spread out over the different
seasons of a career. The burdens on professional societies, universities,
and other organizations lie more heavily.
This paper carries some personal biases of which the reader should be aware.
One is that its focus is not on government or military contexts but on
the private sector, excluding engineering. Another is that, for practical
reasons of my own experience, I am addressing the situation in the United
States.
MAJOR FORCES ACTING ON
MS/OR
Since the destiny of MS/OR is determined significantly by powerful external
forces, those who wish to succeed in this field should be fully aware of
them. Not to be is like trying to fly a plane without understanding the
forces of gravity, lift, thrust, and drag; the results of such ignorance
could be unpleasant--even fatal.
The four forces discussed in this section are the ones I believe to be
of the greatest importance during the last decade or two. The first two--the
microcomputer and communications revolutions--are obvious to even the most
casual observer of the MS/OR scene. The third--the dispersion of MS/OR
in industry--may be the least understood. Finally, I comment on academia's
preoccupation with theory. In each case, I offer an interpretation of the
general consequences for individual professionals, as well as for MS/OR
as a whole.
The Microcomputer Revolution
Everyone is aware of the stunning revolution in small computers (including
workstations) and associated software. Driven by technological progress
in microelectronics and strong market demand--more than 50 million microcomputers
have been sold to date in the U.S.--the microcomputer revolution is expected
to continue apace for some years yet (Hennessy and Jouppi 1991).
Consequences for MS/OR. The impact has been enormous because, in
less than a single decade, the profession's main physical tool has become
ubiquitous and inexpensive. This has many important consequences for MS/OR,
virtually all of them good:
-
The benefit/cost ratio of most MS/OR applications continues to improve
thanks to the decreasing cost of computation.
-
In contrast to prerevolutionary days, there is seldom a computation-cost
barrier to doing MS/OR applications even at the lowest levels of an organization,
or in very small organizations, or in geographically remote locations.
This vastly increases the potential market for MS/OR.
-
Convenient access to microcomputers, the competitive scramble for better
user interfaces, greatly improved software development tools, the object-oriented
paradigm, and other advances have notably shortened development times for
many MS/OR applications. (For nice examples of improved productivity in
applying mathematical programming, see Wagner 1988.)
-
Microcomputers have brought a large measure of freedom from the frustrations
and constraints (organizational, political, and technical) that accompany
having to rely on mainframe computers and data-processing departments.
This, in turn, has triggered a boom in end-user computing.
-
A large and increasing percentage of potential MS/OR clients are personally
familiar with microcomputers, and so tend to be less resistant to computer-based
MS/OR proposals. I believe that this effect, which began only a decade
ago, already has done as much for the acceptability of MS/OR as the roughly
three decades worth of classroom exposure to MS/OR given in management
and engineering programs. Computer applications seem natural to computer
literates, and, in any case, self-contained MS/OR applications on a desktop
machine are far less intimidating psychologically than were the mainframe
applications of yesteryear.
-
MS/OR groups have lost one of their relative advantages, namely, easy
access to computers and mainframe data bases. Ubiquitous computers empower
dispersed MS/OR practitioners and others not identifying with MS/OR, such
as quantitatively-oriented MBAs, who now have the means to apply MS/OR
software with little or no assistance from MS/OR professionals.
-
Affordable high-resolution graphic displays offer new opportunities
for data visualization as a prelude to modeling, and for more realistic
and even animated presentation of results to users (now common in simulation).
-
Microcomputers greatly facilitate communication and demonstration-based
selling for vendors and consultants, both internal and external. Demonstration
diskettes are now the rule for most MS/OR software vendors.
-
The slick user interfaces of the better commercial microcomputer packages
set a standard that client-visible MS/OR application programs must attain,
or risk looking unprofessional by comparison.
-
The decreasing cost of computation is significantly altering the optimal
balance, in research and practice, between mathematical and computational
approaches to problems. For example, it is sometimes almost as easy nowadays
to simulate the transient behavior of a queueing system under realistic
assumptions as it is to use an analytical solution describing steady-state
behavior under unrealistic assumptions. Efron and Tibshirani (1991) lucidly
explain this general idea in the context of statistics.
Consequences for the MS/OR Professional. There are two important
ones to consider:
-
You have a major handicap if you are not knowledgeable about microcomputers:
basic architectures, operating systems, software development tools, and
leading application packages. This knowledge can be acquired through courses,
workshops, and independent study. Do not overlook the newer workstations,
whose price/performance ratio is improving so fast that many professionals
grossly underestimate their potential.
-
Make a habit of thinking about the possible uses of microcomputers
for significant tasks as they arise, keeping in mind that new technologies
typically are used first to facilitate work as it has been done traditionally,
and only later to depart from tradition in creative ways.
The Communications Revolution
The communications revolution comprises such developments as cellular radio,
digital switching, fax, fiber optics, inexpensive micro/mainframe links,
multimedia, network technology (local area, wide area, client-server, etc.),
and satellite transponders and navigational aids. These technologies spawn
distributed data bases, EDI (electronic data interchange), electronic markets,
E-mail and voice mail, enterprise networking, on-line information systems,
and innumerable other applications. An information infrastructure is rapidly
coming into being that will rival the importance of existing infrastructures
like the national highway, telephone, and electric power systems (Hopper
1990, Dertouzos 1991).
Advances in communications receive almost as much press as advances in
microcomputers, and may have equally profound implications. The two revolutions
go hand-in-hand and are occurring for similar reasons. Of their relationship,
Peter Keen says, "We're moving from computing with telecommunications as
an add-on to networking into which computing fits" (Keen 1991).
For convenience, I include under this heading much of what is sometimes
known as the "data explosion" associated with the mass deployment of scanners,
sensors, and other data acquisition equipment, proliferating data base
applications, rapid progress in high-density storage media, and the burgeoning
information services industry. To illustrate, Little (1991) points out
that the introduction of universal product code scanners increased the
amount of data available to consumer-packaged-goods marketing analysts
by two to three orders of magnitude. For another example, Pfizer receives
about 50 million records/month on pharmacies, most of which now are equipped
with computer terminals (Arvesen 1991). As CONDOR (1988, p. 620) puts it,
"Data availability, once the Achilles heel of OR, is now a driving force."
Consequences for MS/OR. There are several important ones to consider:
-
A network of workstations is more cost effective than a mainframe computer
in many applications. Moreover, since data development has always been
an expensive and time-consuming activity in most MS/OR projects, more plentiful
data, and cheaper and faster data acquisition and transmission, are a boon.
Thus, like the microcomputer revolution, the communications revolution
is improving the benefit/cost ratios of many MS/OR applications. The impact
of the new economics is evident at one of the largest MS/OR consulting
firms, which implements nearly all of its new applications on a network
of workstations (Cook 1991). They do this for two reasons: first, to save
mainframe expenses, and second, to facilitate access to mainframe data
in real-time applications that support local decision making.
-
Communications systems and their applications offer numerous opportunities
to apply MS/OR. For example: coordinated inventory management with dispersed
stocking locations (Cohen et al. 1990); electronic markets and EDI systems
(Malone, Yates and Benjamin 1989, EDI Forum 1991), global manufacturing
and logistics planning (McElroy and Trafton 1991), network design (Nigam
1989), optimal vehicle routing with on-board satellite navigation (Spreitzer
1989); and resource allocation and site selection in telemarketing (Spencer
et al. 1990, Quinn, Andrews, and Parsons 1991). Sometimes the kinds of
questions that arise lead to fundamentally new research topics; a good
example is provided by recent pioneering work on inferring waiting-line
statistics from computer-generated transactional data produced by automatic
teller machines, mobile radio networks, and other kinds of computerized
service systems (Larson 1990).
-
There is an accelerating trend toward enterprise networking (Hopper
1990, Rash 1991), a main goal of which is to deploy small computers and
integrate local area networks sufficiently that everyone in an organization
can have convenient electronic access to everyone else in the entire organization
no matter what their location. The reality today is far short of this goal
where heterogeneous networks are involved, but progress in this direction
makes it easier for MS/OR practitioners to stay in close touch with one
another and with their clients and users, makes it easier to do multi-installation
applications, and facilitates distributed modeling and co-development of
new applications.
-
Network-accessible data sources (financial data, geodata, government
statistics, etc.) provide new opportunities to build parts of some models
more quickly and economically.
-
The communications revolution has led to heightened senior management
awareness of the strategic value of information (Keen 1988). Since MS/OR
can be viewed as being in the business of producing high-grade information,
this creates an opening for MS/OR to capture the attention of senior management
for some types of MS/OR applications.
-
The developments of the communications revolution that call for centralized
administration tend to make an organization's information systems (IS)
function more important, which, in turn, presents opportunities for MS/OR
to the extent that IS and MS/OR cooperate. At the same time, other developments
are empowering people at all levels of an organization, leading to more
decentralized decision making (Malone and Rockart 1991) and thus more opportunity
for dispersed MS/OR professionals.
Consequences for the MS/OR Professional. There are two important
ones:
-
You have a major handicap if you are not familiar with information
technology and its applications as they relate to the communications revolution
and data explosion.
-
Make a habit of thinking about the operational, tactical, and strategic
uses of information technology, and about the possible roles for MS/OR
in these uses. For example, ask yourself what your organization would do
differently if it suddenly had a low-cost infinite bandwidth available
for data transport, or if local/wide-area network boundaries were suddenly
to disappear.
Dispersion of MS/OR in Industry
There is only one term to fairly describe the conduct of MS/OR in industry
today: dispersed. The idea that MS/OR is conducted mostly by organized
groups of MS/OR professionals is a myth. MS/OR is practiced mainly by individuals
in myriad types of staff groups and functional areas.
The most dramatic aspect of dispersion is the long-term net decentralization
and disbanding of MS/OR groups. This began about 25 years ago (Radnor and
Neal 1973), and has been confirmed repeatedly by industry surveys. For
example, Bradbard et al. (1987) found 43% fewer MS/OR groups in 1985 than
in 1975. My recent discussions with industry leaders, recruiters, and others
indicate that decentralization/disbanding finally may have run its course,
as reports of new or growing groups roughly match reports of an opposite
nature.
Decentralization has been interpreted as a sign of maturity and success
by some members of the MS/OR community, while others see it as a sign of
failure. In a special report to the TIMS Council, the Management Science
Roundtable, whose representatives are senior MS/OR leaders from industry,
rated this the most severe among seven problems facing the profession (MSR
1986). My own view is that if decentralization is a "victory," then--at
least until now--it has on balance been Pyrrhic.
A less dramatic aspect of the dispersion of MS/OR in industry is the gradual
diffusion of MS/OR ideas into managerial thinking (Wagner 1988), and into
numerous functions and processes throughout many organizations. Seldom,
however, are these ideas identified explicitly as MS/OR. It is difficult
to know the true extent of this victory, but certainly it is more
genuine than Pyrrhic.
The degree of dispersion, for whatever cause, apparently is much larger
than most people realize. According to a landmark Bureau of Labor Statistics
report (BLS 1990), there are about 55,000 Operations Research Analysts
in this country. From this, it follows (Samuelson 1990) that there are
about 48,000 persons in this category who belong neither to ORSA nor TIMS,
but many of whom may be doing work that we would consider to be MS/OR--or
at least related to this field. It seems highly unlikely that more than
a few thousand belong to an organized MS/OR group. Thus, there may be as
many as an order of magnitude more dispersed practitioners who do MS/OR
work than there are in organized groups.
Causes of MS/OR Dispersion. There are quite a few:
-
Lack of functional responsibility makes an MS/OR group an easy target
for cost cutting in a budget crunch and overly dependent on senior management
champions, all of whom eventually move on.
-
There has been a shift for at least a decade toward leaner staffs in
headquarters, flatter organizational structures, and greater decentralization.
The consequences for MS/OR and other staff specialties like long-range
planning often have been dire. The Appendix provides an example taken from
my correspondence about five years ago with the president of a large company
that was dispersing its central MS/OR group.
-
Traditional MS/OR techniques have long been diffusing into other disciplines
and professions such as actuarial science, applied mathematics, computer
science, finance, industrial and other kinds of engineering, logistics,
marketing, and operations management (PoKempner 1977, Gray 1979). This
makes the capabilities of an MS/OR group less unique and increases the
likelihood of MS/OR ideas being used when needed by individuals or groups
with different professional identifications. Moreover, in this age of good
MS/OR microcomputer software, a similar effect occurs each time a quantitatively-oriented
MBA applies an MS/OR technique without much assistance from an organized
MS/OR group. The impact of this diffusion of traditional techniques on
the MS/OR profession has been evident for many years; for example: "The
well known optimization and simulation tools have now been incorporated
into the analytical tool kits of many who are not Management Scientists.
In my company these others are far larger users of simulation and optimization
than are the professional Management Scientists" (Hoffman 1975). The company
referred to is Standard Oil (Indiana). Is it purely coincidental that its
central MS/OR group--one of the most respected in industry--was subsequently
disbanded?
-
The bandwagon effect is long since over; MS/OR is no longer novel,
and therefore is less likely to capture the senior management attention
needed for group formation. The bandwagon effect moved on to DSS, AI (especially
expert systems), neural networks (which Science 1991 calls a "hot
field"), and other panaceas.
-
Many MS/OR groups failed to achieve sufficient success with truly strategic
problems. As John Little puts it, "Many of the problems worked on by industrial
OR groups were tactical in nature and, as a result, senior management gradually
lost interest in high-level groups" (Little, p. 533). Inadequate progress
on strategic applications was judged by MSR (1986) to be the fourth most
serious of seven problems facing the profession. The reasons cited in that
report for this state of affairs include:
Strategic studies are very difficult; they require more
maturity, lots of trust, political and social skills, and a knack for qualitative
as well as quantitative analysis.
Interdisciplinary teams are best for truly strategic problems.
But our profession has somehow lost this inclination.
Such comments notwithstanding, it is only fair to note that MS/OR has had
its share of strategic impacts (Kirkwood 1990).
-
The computer/communications revolution greatly increases the practicality
of dispersed MS/OR work (as well as the work of organized groups).
Few of the above causes of dispersion have to do with how MS/OR is practiced.
Yet it has been suggested that MS/OR practitioners themselves bear some
responsibility for unwanted dispersion. Alleged reasons include arrogance,
poor listening and communication skills, insufficient interest in the business
and its competitive problems, and ineffectual marketing and packaging of
MS/OR.
Consequences for MS/OR. They are mostly unfortunate; only the first
is an unmitigated benefit:
-
MS/OR is being used even more widely, and in more ways, than one would
surmise from the journals and meetings of ORSA and TIMS.
-
Having more MS/OR specialists in functionally-oriented jobs increases
opportunities to introduce good MS/OR technology that meets recognized
functional needs and is fairly simple. On the other hand, fewer MS/OR groups
means fewer opportunities to introduce improved but relatively complex
MS/OR technology into practice. This is so for two reasons: reduced in-house
technical competence, and disruption of the social network by which practitioners
find out about improved technology (MS/OR groups seem to be much more interested
in the academic/practitioner interface than lone practitioners are).
-
Dispersion tends to exacerbate the often lamented identity problems
of MS/OR: the profession and its contributions are insufficiently well
known to its potential clients, its beneficiaries, and the public at large.
See Teach (1991) for the most recent confirmation of this situation.
-
Dispersion tends to shorten the already short career paths available
to people who want to continue in MS/OR. It also reduces the opportunities
for MS/OR to serve as an incubator for higher-level managers, a function
that some central groups have performed with great success (e.g., Machamer
and Smith 1980, Davidson 1981). This comment from an internal IBM report
is a valid lament: "... management science ... career pathing is complicated
by the dispersal of this talent and its associated identification problems
and by the fact that, when identified, the talent resides in many functional
areas serving a variety of managerial levels, focusing on problems with
a diversity of depth and breadth." Shorter career paths will, in turn,
discourage very able people from entering MS/OR or staying in it.
-
Dispersion is an obstacle to cross-functional applications, which usually
require multitalented MS/OR teams and senior management involvement. This
is a particularly grave consequence because such applications are an increasingly
important area of opportunity for MS/OR. As pointed out in the internal
report quoted in the previous point: "... management science has its greatest
impact when interfunctional problems are to be attacked--and centers of
management science competence are most effective in addressing these problems."
-
Much of the demand for MS/OR practitioners has moved away from those
with only specialized skills and toward those with generalist and functional
skills. Many observers--including the author of The Conference Board's
independent study of MS/OR (PoKempner 1977)--see this as a potentially
fatal threat to maintaining the critical mass of technically qualified
practitioners needed to do advanced applications and sustain the health
and relevance of MS/OR's academic infrastructure. Without such a critical
mass, these observers say, MS/OR is doomed to sterility.
-
A corollary of the previous point is that the educational preparation
of Master's-level practitioners needs to be changed if demand is to be
satisfied. This was identified in MSR (1986) as the number two problem
facing the profession. Another corollary is that the industrial market
for MS/OR Ph.D.s has become more concentrated in industrial research laboratories
and consulting firms.
-
First-generation decentralized practitioners may not regenerate themselves.
For several reasons, it is much easier for an MS/OR person to move into
a dispersed position from an organized MS/OR group than directly from school
at the entry level. It follows that the population of MS/OR people in dispersed
positions will tend to diminish with the attrition of those who got there
by being decentralized.
-
Dispersion tends to cause good people to leave the profession to the
extent that they like to work with other good people with similar interests
(Ranyard 1989).
Consequences for the MS/OR Professional. There are two important
ones:
-
Be realistic about where most industrial MS/OR jobs are: in dispersed
positions rather than in organized MS/OR groups. If you are a practitioner,
this has strong implications for your choice of educational and professional
development investments.
-
Be a perpetual student of the causes of unwanted dispersion, and try
to become invulnerable.
There may be lessons to be learned from one of MS/OR's neighbors, MIS,
which also has been experiencing pressures in the direction of dispersion.
One such pressure arises from the downsizing trend. The most militant
form of downsizing advocates replacing most large computers by networks
of small ones, while a more benign form is content to encourage the migration
of many applications from mainframes and minicomputers to networked microcomputers.
Even the benign form is a threat to centralized MIS. Consider this statement
by the system architecture manager of a manufacturing company:
[Downsizing will] radically affect the end user. Not only
are we changing the hardware, but we're changing our philosophy of what
DP is. Our strategy is to eliminate the central [MIS] staff as we know
it today. There will probably be more programmers than ever, but they will
be working in the end-user departments or at the divisional level, not
at the corporate level. (Ryan 1991, p. 232).
Does this sound familiar? Note that there are also pressures in the direction
of upsizing and centralization (e.g., von Simson 1990), some of which may
have parallels for MS/OR.
The Theory-Oriented Academic Reward Structure
Two facts conspire to create a powerful force on the academic development
of MS/OR. First, a sufficient number of publications in top journals like
those of ORSA and TIMS is a necessary condition for appointment and promotion
at nearly all research-oriented schools. Second, theoretical contribution
has come to be the most reliable publishability criterion by far for the
top journals. It follows that Ph.D. students and young academics at research-oriented
schools feel obliged to make theoretical research their top priority, and
they do.
Given this situation, and the fact that teaching is the only other appointment/promotion
criterion that can approach research in importance, is it any wonder that
few academics choose to make a priority of integrative or applied work?
Yet widespread, successful practical application of MS/OR is the only real
justification for academic MS/OR programs in the first place.
Next to the stimulation provided by real applications, I consider this
force to have been the dominant one shaping the evolution of MS/OR in academia,
just as the previous force (dispersal) has been the dominant one outside
of academia.
This force is not unique to MS/OR. Many of the nation's research universities
exhibit similar symptoms to a greater or lesser extent in many other fields.
There has been active discussion at some institutions of possible remedial
actions. My own university, for example, has just launched a university-wide
debate on this and related issues, which have been nicely crystallized
in Pister et al. (1991).
Consequences for MS/OR. There are a number that are important:
-
MS/OR now has an awesome theoretical base, and there is a plentiful
supply of fine theoreticians.
-
The preoccupation of academia with theory sometimes results in excellent
work that strengthens the foundations of MS/OR for decades to come (witness
most of the contributions recognized by the Lanchester and von Neumann
prizes). But more often, the result is work that is highly incremental
or so specialized that eventual application is unlikely. In the words of
a recent review panel (ORSA 1991), "... much of what is published is apparently
never again applied or built upon. This results in wasted pages and spurious
signals to promotion committees, and it fogs the image of our field to
outsiders." Peter Keen, not known for veiling his views, puts it more strongly:
The theory is largely weak, the research is mediocre and
far too confined in themes and methods, and most of the publications are
mediocre. ... Ten years ago, MS owned both the intellectual and technical
base of the field. It has lost the technical base and has seen the intellectual
base trivialized and disregarded. The key for me is re-establishing the
message that MS represents a powerful approach to problem definition and
problem solving, showing that that approach is targeted to relevant problems,
and demonstrating that MS professionals have the needed attitudes, listening
skills, and respect for real decision makers needed to have an impact (Keen
1991).
In other words, rigor mania can lead to rigor mortis when relevance is
neglected.
-
MS/OR practitioners commonly complain that academia is not doing the
kinds of things that could help them. For example, a preeminent practitioner
who used to be an academic has said, "Too often, academicians are encouraged
to spend their time on meaningless research by an outmoded tenure and reward
system" (Cook 1990, p. 27). Such feelings may be widespread. One indication
of this is that many practitioners judge even the leading MS/OR journals
to be of little or no use (Bradbard et al.), although Interfaces is
a conspicuous exception.
-
Academic programs in MS/OR tend to become narrower and more specialized.
Constantly expanding technical topics tend to crowd out studies of business
and government, to the probable detriment of preparation for applied work.
Many academics join the practitioners in bemoaning the direction taken
by academia over the last two decades or so. I have been among them on
occasion. But such laments do not cut to the heart of the matter, which
I believe to be this: there is a fundamental mismatch between supply
and demand for MS/OR technology. Academia, largely decoupled from the
economic fortunes of its practitioner constituency, has forged ahead and
produced so far beyond the real-world demand for its intellectual product
that its fruit appears to many to lack relevance. This situation is exacerbated
by the dispersion of MS/OR in industry, which tends to obscure the visibility
of most applications and bias them toward low technology.
How could this mismatch be reduced? The only obvious remedial steps are
those aimed at: a) downsizing graduate academic programs in MS/OR, b) transforming
the reward structure at some universities into one more friendly to integrative
and applied work, and c) increasing the real-world demand for MS/OR technology.
The third main section of this paper focuses on the last of these.
It bears pointing out that there may be a slightly devious way to accomplish
much the same effect as b) without actually requiring universities to change
at all: ORSA and TIMS could simply tilt their publication criteria in the
direction of integrative and applied work. Additionally, the local culture
of most academic MS/OR units can be changed in this direction if most of
the influential members agree that this is appropriate and then live
that belief.
Consequences for the MS/OR Professional. There are three important
ones:
-
If you are a student, strive to balance the time you spend learning
technical subjects and the time you spend learning about application contexts.
-
If you are in academia, spend some time doing challenging consulting.
Do it for the stimulation, not the money. This experience will confer many
benefits on your research and teaching, and will help to promote the health
of your chosen profession. The proportion of time you spend doing teaching,
research, consulting, and professional service can shift over the years,
but do not wait too long before getting some consulting experience; it
will stimulate new ideas and lead you to worthwhile research problems.
Moreover, your publication record should benefit rather than suffer so
long as you avoid routine consulting. A second and perhaps equally important
step that you can take is to give more attention to professional practice
in the curriculum, including practical student projects in some courses
and summer internships (Miser 1976, Bajgier et al. 1991).
-
If you are a practitioner, allocate part of your time to interactions
with academia. Consider providing some incentives to academics to do applied
work as a counterweight to the usual academic incentives.
MAJOR TRENDS
Of the important trends evident today in MS/OR (but only partly in response
to the causative forces discussed in the previous section), this section
discusses three: the rapid dissemination of MS/OR tools, the declining
number of native-born students in MS/OR graduate programs, and the persisting
management apathy toward MS/OR.
For each of these trends this section interprets the consequences, both
for the profession as a whole and for the individual MS/OR worker.
Rapidly Disseminating MS/OR Tools
One of the consequences of the computer/communications revolution is an
explosion of commercial activity in MS/OR-related software. It amounts
to a massive new distribution channel for MS/OR technology. Consider:
-
Tens of millions of spreadsheet installations exist, each offering
quite general modeling capability. Spreadsheets have some important limitations
for MS/OR applications, such as excessive focus on model instances to the
exclusion of model classes, a limited range of mathematical solvers, poor
maintainability, and awkward extensibility. But some of these limitations
are gradually being lifted, and certain capabilities with great potential
are emerging, such as automatic links between separate spreadsheets and
between spreadsheets and powerful data base packages. By the end of 1991,
Frontline Systems had shipped over a million copies of Lasdon and Waren's
GRG2 linear and nonlinear optimization code modified to work with the leading
spreadsheet packages. The version shipped also has rudimentary integer
programming capability.
-
LP packages, including some very good ones, have become a widely and
inexpensively available "commodity." Llewellyn and Sharda (1990) list 37
packages for PC compatibles, and several more for workstations.
-
Graphically oriented simulation packages have become widely available
for microcomputers (e.g., Banks et al. 1991, Swain 1991). At one consulting
firm, such packages came into such widespread use among their clients that
some resources could be reallocated from basic to more advanced simulation
support and to other consulting areas that have not yet achieved such wide
client acceptance (Lucas 1991).
-
Many nice microcomputer packages have become available for data visualization,
decision analysis, expert systems, forecasting, Monte Carlo simulation,
multi-attribute choice over discrete alternatives, nonlinear optimization,
project management, and data analysis/statistics. Announcements and reviews
of such packages appear regularly in the European Journal of Operational
Research, OR/MS Today, personal computer magazines, and elsewhere.
-
A recent survey of software for logistics lists numerous microcomputer
packages with modeling capability, including 90 for inventory planning
and forecasting, 87 for traffic routing and scheduling, and 56 for physical
distribution system modeling (CLM 1991a).
These observations show that, whether or not MS/OR professionals like it,
give permission, or actively assist, many MS/OR tools are widely available
to anyone with a modest budget. Fylstra (1991) calls this the "democratization
of operations research." Others view it as vulgarization.
A companion trend is evident in the diffusion of MS/OR techniques into
other disciplines and professions, as noted previously (e.g., actuarial
science, applied mathematics, computer science, finance, industrial and
other kinds of engineering, IS, logistics, marketing, and operations management).
MSR (1986) discussed the emergence of a mass market in MS/OR-related software.
The 30 participants in that panel were about equally divided between viewing
it as a "problem" versus as an "opportunity." A typical view from the first
camp:
Another aspect of the problem: spreadsheet packages cheapen
our profession because untrained people claim to be doing "modeling". Most
managers can't tell the difference, and so judge what MS/OR offers by those
rudimentary and questionable examples.
A typical view from the second camp:
There are advantages as well as dangers: we are relieved
of routine applications; clients are better prepared as a result when they
finally do come to a MS/OR specialist; and some users of these packages
may be attracted to deeper involvement with MS/OR.
A third participant offered this constructive suggestion:
We need to devise "migration paths" from simple, standard
modeling applications to more complex ones. This begins with making sure
that MS/OR software users a) know where to turn for help in their organizations,
and b) understand the value added potential of MS/OR expertise.
There has been only limited discussion to date in MS/OR publications of
the professional implications of the rapid commercialization of MS/OR software.
For example, veteran observer Saul Gass urges steps to inform untrained
users about the importance of always using these tools in the context of
the model life-cycle process that he has articulated so well (Gass 1990).
My personal view is that, on balance, this trend is a good one. It enables
our technology to have a much greater impact on the world and frees our
energies to pioneer new technologies and new applications. There are indeed
associated risks--poorly conceived and documented models, excessive emphasis
on ease of use at the expense of functionality and deeply analytic thinking,
loss of control over data used for decision making, and so on--but I see
these as challenges to educators, trainers, and software designers that
can be overcome through hard work. In any case, we have no responsible
choice but to do what is necessary to ensure that this irreversible trend
works to our advantage.
Consequences for MS/OR. There are many:
-
Every organization, even the smallest, can now afford MS/OR software
tools.
-
The MS/OR community no longer has a monopoly on its technology, and
so must share it gracefully or be bypassed.
-
Among other things, graceful sharing means teaching nonspecialists
how to evaluate commercially available MS/OR tools, and how to use them
effectively. This mission seems appropriate for related non-MS/OR curricula
as well, reaching all the way down to high school, where computerized models
are increasingly used to motivate mathematics education.
-
Huge numbers of people untrained in MS/OR are gaining access to its
tools for the first time, and will a) have failures that will give MS/OR
a bad name, b) have successes that will give MS/OR a good name, and c)
become candidates for more serious MS/OR assistance at a later date.
-
Some managers interpret the ready availability of MS/OR software as
lessening the need for staff specialists in MS/OR. It would be prudent
to marshal the arguments for why exactly the opposite is true.
-
The profession's visibility/identity problem is exacerbated to the
extent that disseminated tools are not identified with MS/OR, and are used
by people from other specialties.
-
The overemphasis of much MS/OR-related software on flashy user interfaces
and ease of use tends to trivialize MS/OR's analytical approach. We need
to combat this tendency.
-
Microcomputer tools for MS/OR offer a practical way to enthrall and
empower students in virtually all educational programs with an MS/OR component,
from undergraduate through Ph.D. programs.
-
There is a growing demand among MBA students for courses on how to
evaluate and properly use microcomputer tools for decision analysis, optimization,
project management, simulation, advanced spreadsheet modeling, and statistics.
This demand is not limited to those whose interests are directed toward
MS/OR.
Consequences for the MS/OR Professional. There are three important
ones to note here:
-
Allocate some time to keeping up with the rapidly developing MS/OR-related
software scene through magazines, review services, and other sources. Become
familiar with a leading package in most of the major MS/OR-related software
categories, and try to understand its proper uses and limitations within
the context of established MS/OR theory. Retrain yourself periodically
as the state of the art moves forward. (Journal editors take note: there
is, at present, a severe lack of substantive MS/OR-related software reviews.)
-
In addition to traditional publications, consider commercial software
as an outlet for disseminating your contributions to MS/OR technology.
-
Accept the fact that this trend places a large burden on our community
to help nonspecialists make proper use of MS/OR-related software and understand
that there is a lot more to MS/OR than is evident in this software. Do
your part to help.
Vanishing Natives
The title of this subsection is taken from a recent special section on
science careers in Science, which identifies the shift from American
to foreign-born science students and practicing scientists in the U.S.
as one of the top 10 trends pertinent to career choice (Science 1991).
Certainly this trend is also a strong one in most MS/OR graduate programs
in the U.S.; John White pointed out in last year's Omega Rho Lecture that
"the number of U.S.-born graduates from MS/OR programs has steadily declined
over the past decade" (White 1991, p. 184). MS/OR-specific statistics are
not available, but the situation probably is similar to that in engineering
and mathematics, where foreign citizens now earn about half of all Ph.D.
degrees (NRC 1989, Pool 1990).
White points out that the influx of international students has been a blessing
to MS/OR graduate programs, but betrays deficiencies in the U.S. educational
system and places MS/OR in a vulnerable position with respect to its future
supply of human resources.
The causes of this trend include the following:
-
A declining U.S. educational system, competition from other exciting
career alternatives, and perceptions of a poor MS/OR job market diminish
the number of Americans who are prepared and motivated to pursue graduate
education in MS/OR. See NRC (1989) for a discussion of these and other
causal factors in the neighboring field of mathematics, where the absolute
number of Ph.D. degrees awarded to U.S. citizens has fallen by about half
since 1970. The loss of American students would have led to moribundity
for many MS/OR programs were it not for the influx of large numbers of
well prepared foreign-born students.
-
The U.S. is very attractive to top students from other countries: graduate
MS/OR education here is as good as can be found anywhere in the world (especially
at the Ph.D. level), U.S. degrees carry a premium in many countries, and
the U.S. is still seen as a land of opportunity, especially in technical
fields (which immigrants perceive to prize competence over communication
skills).
These causes will be exacerbated during the 1990s by demographic factors.
The well known baby boom maximized the number of citizens of graduate school
age during the mid-1980s. The trend will be strongly downward through the
1990s before beginning to pick up again around the turn of the century.
With a peak-to-valley ratio in excess of 1.4, this is not a small effect.
Consequences for MS/OR. The effects of this trend are mixed:
-
Pending reversal of the first cause, MS/OR graduate programs must either
downsize or maintain their populations by increasing the proportions of
international students.
-
Industry recruiters find it difficult to meet their needs from an increasingly
foreign-born supply. There are three reasons for this. First, foreign-born
Master of Science students usually need to be permanent residents, but
few are. (More employers are willing to sponsor Ph.D. students.) Second,
strong communication skills are at or near the top of most recruiters'
list of requirements (Buckholz 1990, Hoffman 1991), but the proportion
of foreign-born applicants who meet this requirement is considerably smaller
than for U.S.-born candidates owing to language and (sometimes) cultural
differences. Third, the emphasis that many international students place
on technical studies necessarily exacts a price in terms of interest and
preparation in business and the particulars of application domains.
-
Foreign-born MS/OR professionals, with their language and cross-cultural
skills, are a key resource for many MS/OR opportunities arising from the
powerful trend toward economic globalization (Reich 1991).
-
If our profession does not take vigorous steps to assure an ample supply
of qualified human resources, then its future vitality will be jeopardized
and it may be unable to exploit great opportunities like those to be discussed
shortly. At least two remedial directions seem promising.
a. Correct inappropriately pessimistic perceptions concerning
the MS/OR job market. For example, we know from the first section that
perceptions based on the plight of organized industrial MS/OR groups are
out of touch with the current reality of dispersed employment. Seldom is
it easy to change long-standing perceptions, but for once MS/OR possesses
something close to a quick fix in the form of an authoritative government
report (BLS 1990) that pronounces "operations research analysts" the 10th
fastest growing among all occupations in the 1990s! Faster than
accounting, computer programming, engineering, financial management, health
services management, management consulting, medicine, and more than 300
other occupations. Moreover, unpublished supporting tables (obtained for
me by D. Samuelson) explain where that growth is expected to occur. This
conclusion is also found in the Occupational Outlook Quarterly (OOQ
1990). These two publications deserve vigorous publicity because job-market
perceptions influence graduate-student applications.
b. Make the attractions of MS/OR better known to U.S. high
school students and undergraduates. The ORSA videotape is a good step in
this direction. Cooperation with the mathematical and scientific communities
may be fruitful, as they have an established presence throughout the educational
system and have succeeded in mobilizing a high level of visibility for
the need to improve their infrastructure (e.g., NRC 1989).
Concerning the second point, I believe that overtures to the mathematical
and scientific communities would be met warmly. One reason is that the
(complete) list of topical areas "deserving greater emphasis" in NRC (1989)
are: probability, exploratory data analysis, model-building, operations
research, and discrete mathematics. And the widely noted report "Renewing
U.S. Mathematics" (BMS 1990) explicitly defines statistics and operations
research as part of the mathematical sciences, and includes the following
topics on its list of 27 "recent research accomplishments and related opportunities":
statistical methods for quality and productivity, mathematical economics,
parallel algorithms, randomized algorithms, interior point methods for
linear programming, and stochastic linear programming.
Consequences for the MS/OR Professional. Two important ones may
be discerned:
-
If you are foreign-born, it is vitally important to perfect your language
and other communication skills, especially if you are or wish to be a practitioner.
This applies not only for entry, but also for advancement. You may take
some consolation in the fact that many--perhaps most--native English speakers
are poor communicators, especially on paper.
-
If you are foreign-born, consider international work, where your background
is a special advantage.
Persisting Management Apathy Toward MS/OR
Most potential managerial clients have been exposed to MS/OR by now: CPMS's
Edelman Competition has been uncovering and publicizing success stories
for two decades; quantitative methods and statistics courses have been
almost universally required by MBA programs for about three decades; and
the industrial history of MS/OR is, for practical purposes, in its fifth
decade. Given these facts and the obvious importance of MS/OR to management,
it is natural to expect a high degree of management enthusiasm by now for
our profession. Yet management is only marginally less indifferent to MS/OR
today than it was decades ago. There are numerous notable exceptions, but
only a small proportion of all senior managers recognize the wide applicability
of MS/OR within their organizations and insist on strengthening their MS/OR
infrastructure.
In other words, the trend is that there is little or no trend in
the direction of diminished management apathy toward MS/OR. MSR (1986)
rated this the number three problem (out of seven) facing our profession.
It has numerous causes:
-
The root cause, in my opinion, is that it is not the business of managers
to be concerned with the power or potential of any discipline--whether
it is MS/OR, economics, psychology, statistics, or any other field with
important managerial applications. Rather, it is the business of managers
to be concerned with the management of their businesses. Period.
-
It is well known that the MS/OR approach tends to clash with the personal
style of senior managers (e.g., Mintzberg 1973, Hammond 1974, Lilien 1987).
Moreover, conventional wisdom has it that a personal dislike of mathematics
and technical details is one of the reasons why people choose to become
managers.
-
Many MS/OR professionals communicate poorly. As several participants
in MSR (1986) put it, "We have not communicated very well with senior managers.
For example, we use too much jargon and do not know how to say crisply
what we can do or have done."
-
MS/OR has a visibility problem. The popular and business press largely
ignore MS/OR, sadly even the great success stories. Thus, the management
community receives very little external confirmation of MS/OR's worth in
the normal course of their reading.
-
In some cases, managers fall prey to misunderstandings about MS/OR.
One is the belief that the increased complexity of today's problems renders
them less amenable to modeling and the simplistic assumptions commonly
associated with modeling (independence, linearity, normality, stationarity,
steady state, etc.). Another is the belief that the rapidly falling cost
of computation now enables most problems to be overwhelmed with computer
power, thereby rendering MS/OR's elegant approaches less essential than
previously. BMS (1990) mentions a similar misunderstanding in the context
of the need for mathematics.
I conclude from the first two causes that management will continue indefinitely
to be more apathetic toward MS/OR, on average, than we would like. Moreover,
I cannot fault the rationality of this posture. Of course, this is not
to deny the temporary bursts of interest that can occur in the wake of
rare events--like the one triggered by N. Karmarkar--which happen to be
picked up by the mass media.
To be fair, I must note that there are astute observers of the MS/OR scene
who take a somewhat more sanguine view. Among the most articulate of these
is Harvey Wagner, who believes that:
... the concepts and vocabulary of operations research have
become a pervasive part of the thinking of modern American industrial managers
... many executives throughout the typical large business organization
understand the underlying concepts of formal model analysis and make constant
use of this language. ... The progress to watch for in the next 10 years
... will be evident in the growing and continuing successes of general
managers who are making consistent and effective strategic use of formal
operations research models (Wagner, pp 797,803).
Consequences for MS/OR. There are at least two:
-
The MS/OR community can stop hoping for a major management demand pull
to materialize. We must accept responsibility for creating our own demand,
as by becoming indispensable for important organizational functions or
by doing the kinds of things mentioned in the next section.
-
Educators should consider intensifying their efforts to provide a better
appreciation of MS/OR to students who may be future clients. For example,
they could integrate MS/OR material into more non-MS/OR courses in MBA
programs, especially the functionally oriented ones like accounting, finance,
and marketing. (This is a thankless task at many schools, however, so new
incentives will be needed before much more of this will be done.) They
could be more aggressive about exposing students to Edelman finalist papers
and other outstanding applications. And they could specifically target
common management misunderstandings like the ones mentioned in the last
cause listed above.
Consequences for the MS/OR Professional. The most important one
is to learn how to market MS/OR's capabilities in the context of organizational
needs. This involves pondering what MS/OR has to offer in various situations,
collecting MS/OR success stories, being a student of organizational behavior,
and studying the ways of master practitioners and teachers.
This completes my survey of the main forces and trends as I see them. They
provide an essential context for understanding MS/OR's opportunities, to
which I now turn.
OPPORTUNITIES
Of the major opportunities facing MS/OR today, I have selected five for
recommendation here: ride the computer/communications revolution, support
dispersed practitioners, focus on the service sector, stress embedded applications,
and go into the quality business. None is in conflict with any of the major
forces and trends discussed in the preceding two sections, and most draw
strength from them.
The orientation adopted here slants toward the practice of MS/OR because
that is the fountainhead of the profession's future vitality. See CONDOR
(1988) for a discussion more concerned with theoretical research opportunities.
Ride the Computer/Communications Revolution
Because microcomputers and communications facilities are among the main
physical tools of MS/OR, and because MS/OR provides so many opportunities
to make good use of them, the computer/communications revolution is an
exceedingly powerful engine for the progress and acceptance of MS/OR.
Taking advantage of this engine requires MS/OR professionals to a) maintain
an up-to-date understanding of computer/communications technologies, b)
be imaginative and aggressive about opportunities to apply these technologies,
both to MS/OR work proper and to organizations, and c) become closely identified
in the minds of potential MS/OR clients with innovative and truly productive
uses of computer/communications technology, and with the ability to add
value to all sorts of data through model-based analysis. These are, to
some extent, a natural by-product of normal MS/OR activity, but I am proposing
that we do these things to a greater degree, and more deliberately, than
would be the case if we were not consciously trying to take full advantage
of the computer/communications revolution.
The last two require a generous measure of inspiration. It is difficult
to plan for inspiration, but there do exist plannable activities that can
lay the groundwork for success and perhaps even stimulate inspiration to
some extent. These include:
-
Establishing an evaluation, recommendation, and training program for
useful commercial MS/OR software suitable for end users in such areas as
decision analysis, optimization, project management, simulation, advanced
spreadsheet modeling, and statistics. This is necessary anyway in response
to the rapid dissemination of MS/OR tools as discussed earlier, and it
helps acquaint practitioners with the current concerns of end users.
-
Promoting advanced workstations and enterprise networking (as discussed
in the first section's treatment of the communications revolution).
-
Periodically scanning new information technologies for applicability,
preferably with an embedded MS/OR component.
-
Offering to help re-engineer business processes, especially when this
involves significant amounts of information technology (Hammer 1990).
-
Helping Information Systems people assimilate better decision technology.
IS needs the analytical approach of MS/OR to turn data into information
and information into the knowledge and insights required for improved decision
making; IS and MS/OR each have much to gain from close working relations,
both in academia and in industry.
Riding the computer/communications revolution was by far the number one
opportunity identified in MSR (1986), and is so widely appreciated that
there is little need to expand upon it here.
Support Dispersed Practitioners
As we saw in the first section, most MS/OR practitioners today are dispersed
throughout their organizations. They work in engineering, finance, information
systems, logistics, manufacturing, marketing, planning, scheduling, services,
and many other areas. Yet it seems that only a small percentage of these
practitioners--perhaps as few as 10%--ever come to ORSA/TIMS meetings or
read the societies' journals; the rest are, for all practical purposes,
isolated from the mainstream MS/OR community.
This isolation is at once a great misfortune and a great opportunity for
our profession. I believe that dispersed practitioners would become significantly
more effective if they could easily network with one another and with the
mainstream, have access to the news and literature of the field, and attend
professional development events from time to time. This increased effectiveness
should, in turn, translate into much more of the one thing that controls
the ultimate destiny of MS/OR: positive impact on the real world. Therefore,
the MS/OR community and its infrastructure would be well advised to reorient
itself much more toward the needs of dispersed practitioners.
The Role of the Societies. ORSA and TIMS have an important role
to play in catalyzing such a reorientation. More than that, I would say
that they cannot pronounce themselves truly successful professional societies
until they find a way to represent the interests of more than a tiny minority
of all professionals doing work related to MS/OR. This would seem to require
that the societies a) sponsor an in-depth study to confirm the estimates
of BLS (1990), and to probe the reasons for and consequences of decentralization
and the other mechanisms of dispersion, b) find out how the needs of dispersed
practitioners differ from those in organized groups, c) redesign existing
ORSA and TIMS services and create new ones to better meet these needs,
and d) undertake an ambitious membership drive among dispersed practitioners.
Extensive consultation with dispersed practitioners themselves will be
necessary at each step. Such a project has the potential to be epoch-making,
and would offer the possibility of greatly increasing ORSA and TIMS membership
for the first time in 20 years.
If this project is undertaken, it would be wise to cast it in a broader
perspective by simultaneously studying the total MS/OR job market in detail.
Presently the job market is poorly understood, partly because of the effects
of dispersion, with no comprehensive, generally available statistics or
interpretive analyses beyond the remarkable Bureau of Labor Statistics
report mentioned earlier (BLS 1990). A first step might be to segment the
job market into submarkets. Employment data could be gathered, analyzed
for trends and career patterns by submarket, and made available publicly
in summary form on an annual basis. This would be a valuable service to
graduating students, professionals seeking better work, employers, and
educational program designers. Some neighboring fields, like computer science,
have had such a service for many years (Gries and Marsh 1992; see also
Jarvenpaa, Ives and Davis 1991).
Not only is the need great for ORSA and TIMS to support dispersed practitioners
in new ways, but the communications revolution makes this mission much
less daunting than it would have been in the past. E-mail, enterprise networking,
fax, voice mail, and other capabilities make it easier every year for dispersed
practitioners to be in touch with counterparts elsewhere in their organization,
and with their users and clients. Their work style is becoming more communications
oriented.
Consequently, the time may be right to launch a bold new service specifically
for dispersed practitioners: a national network that aims to help them
become more effective in their jobs, and to make them an integral part
of the MS/OR community from which they are now largely estranged. Participating
in a well designed network should be accepted as a natural step by many
of them, and ought not to be technically difficult in most cases.
ORSA and TIMS could design and operate such a network. It could offer a
variety of useful services such as direct contact with others in all walks
of the profession, downloadable software (commercial and otherwise), focused
discussion groups, software reviews and surveys, information sources, interactive
magazines, notices of meetings, and on-line newsletters and reports. Since
dispersed practitioners tend to be strongly focused on the needs of their
organizations, it may be desirable for some services to be industry-specific.
The Role of Universities. To support dispersed practitioners, an
important step would be to increase the amount of functional training in
MS/OR degree programs. For example, schools of management could increase
the number of MBA courses in their Master of Science programs at the expense,
presumably, of more specialized training. More generally, to the extent
that they wish to train dispersed practitioners, MS/OR curricula also need
to develop more realistic teaching materials, develop links to other curricula
whose graduates can benefit from MS/OR, stay abreast of rapidly changing
industrial technology, stress fundamentals and the techniques most valuable
in practice, give serious attention to every step of the modeling life-cycle,
teach how to evaluate and properly use the kinds of commercial MS/OR software
on which dispersed practitioners largely depend, and give more attention
to the interpersonal and other skills needed to succeed as practitioners.
Another useful step would be for universities to offer professional development
seminars and courses specifically for dispersed practitioners, possibly
on an industry-specific basis.
In summary, I believe that reorienting the academic and professional society
components of MS/OR's infrastructure toward the needs of dispersed practitioners
has the potential to profoundly improve the evolutionary path of our profession.
Focus On the Service Sector
It is well known that the U.S. economy has been restructuring itself for
several decades toward services at the expense of goods production. Industries
like automobiles, petroleum, primary metals, and textiles have been shrinking
while banking, business services, health services, insurance, public utilities,
transportation, and other service industries have been expanding.
Growth of the Service Sector. In the early 1950s, these two sectors
had about equal shares of the gross national product. By 1974, services
were half again as large as goods production, and by 1986 it was twice
as large. The trend continues, with services today about $3 trillion/year.
Moreover, a similar shift has been occurring in other industrialized economies.
(In all cases, I exclude the government component of GNP and take my figures
from Bush 1991.)
The emergence of the service sector looks roughly the same in terms of
nongovernment employment. In 1988 there were 2.2 people employed in services
for every person employed producing goods, and the ratio is forecast to
increase to 2.7 by the year 2000 (BLS 1990). If government employment is
included in services, the ratios become 2.8 and 3.3.
The new dominance of the service sector is not only a national economic
fact. It is also a reality for our profession: the Bureau of Labor Statistics
forecasts 6 new jobs for operations research analysts in this sector for
every one in the goods-producing sector between 1988 and 2000 (BLS 1990)!
MS/OR's Response. Has the MS/OR community been giving an adequate
share of its attention to services in recent years? My attempts to answer
this question suggest a curious schism within the profession: service seems
to be grossly underrepresented relative to goods-production in academia,
but it seems to be only modestly underrepresented if at all in industry.
Some evidence relating to academia:
-
During the last 3 years, the orientation of submissions to Operations
Research has run roughly 6:1, goods-production over services. Comparable
statistics are not available for Management Science, but my inquiries
there give no cause to hope that the balance is any better.
-
A survey of MBA courses at 20 leading schools (Carraway and Freeland
1989) found that 90% of the schools have an elective course in production
and inventory systems, whereas only half that number have one in service
systems; yet service is much more popular as an elective than production.
Material on service was found to have increased during the 1980s in required
operations management courses, but the amount of attention devoted to this
topic is still small compared with more traditional topics related to production.
-
The most recent (November 1991) ORSA/TIMS Joint National Meeting had
more than twice as many sessions clearly relating to production as sessions
relating to service. I assume that this is typical for these semiannual
events, which mainly cater to academia.
In addition, it is my impression of MS/OR textbooks that the goods-producing
sector, rather than services, provides the dominant context for the material
presented.
Quite a different picture emerges when one examines industry:
-
The finalist papers of recent Edelman competitions (the last four)
favor service over production, 18 to 10. This suggests that the service
sector may be receiving its just share of high-impact MS/OR applications.
-
According to the U.S. Department of Labor, the number of Operations
Research Analysts employed in the service sector in 1988 was about 2.9
times as large as in the goods-producing sector (BLS 1990). Moreover, this
ratio is forecast to rise to 3.7 by the year 2000 because the expected
employment growth rate for operations research analysts is twice as high
in the service sector. These ratios are significantly larger than what
one would expect based on a pro rata share of total employment in the two
sectors, which indicates that MS/OR's penetration is highest in the service
sector.
-
Morgan (1989) reports that none of the 15 industry surveys reviewed
provide any conclusive evidence that the service sector uses MS/OR proportionally
more or less than the goods-producing sector does. (An earlier survey by
The Conference Board--PoKempner 1977--did find a bias: 40% of the responding
companies in the production sector had MS/OR groups, whereas only 29% of
those in the service sector did.)
I conclude from this and other evidence that, although the service sector
deserves about twice as much of attention as the goods-producing sector
does, it is getting nowhere near that level of attention in academia. The
situation looks considerably better outside of academia, but there is still
room for improvement.
MS/OR's Opportunity. Could it be that lack of opportunity accounts
for MS/OR's underemphasis on research, teaching, and possibly practice
in the service sector? I seriously doubt it, based on what I know of the
service industries and the high-impact work reported in recent Edelman
finalist papers. To say nothing of opportunities at all levels of government,
for which much greater efficiency is essential if high expectations for
public services are ever to be reconciled with structural budgetary woes.
Others in a position to know agree (e.g., Larson 1988).
A tangible indication of MS/OR opportunities in the service sector is the
fact that, in the aggregate, productivity has stagnated there since the
early 1970s while productivity in the goods-producing sector has continued
to rise steadily (Harker 1991, Roach 1991). This is in spite of the service
sector's massive investment in information technology (IT): it now owns
about 85% of the installed IT base (Roach). Clearly there is a ripe opportunity
for MS/OR to boost productivity, perhaps partly by helping to put all that
IT to better use. Roach argues that service organizations need better measures
of productivity and service quality, and improved activity-based managerial
accounting, before proper remedies for productivity stagnation can be found.
To an MS/OR person, this is a clarion call for modeling. See Harker for
further discussion of MS/OR's potential to improve productivity in the
service sector.
Another indication of opportunities for MS/OR in the service sector has
just surfaced in a book-length study done for the Council of Logistics
Management (CLM 1991b). This study argues that logistics, broadly defined,
is as critical to success in the service sector as it is in the goods-producing
sector. Moreover, the study concludes that logistics techniques and skills
are highly transferable between these sectors, and yet have barely begun
to be used. Although I must discount the very last assertion, the study
nevertheless poses a provocative challenge to the MS/OR community.
The study estimates that the service sector spends about 75% of its operating
costs on logistics-related processes. It observes that, if logistics improvements
in service can reach the level achieved in goods-production during the
last decade, the savings potential is about $200 billion per year. A significant
part of that savings potential could be within reach of MS/OR.
To understand the role of MS/OR in helping to achieve this potential, it
is necessary to distinguish between supply chain logistics and service
response logistics. The former has to do with the acquisition and distribution
of goods: purchasing, transportation, inventory control, materials handling,
manufacturing, distribution, and related functions. The latter has to do
with "the process of coordinating nonmaterial activities necessary to the
fulfillment of the service in a cost- and customer service-effective way"
(CLM 1991b): anticipating customer needs, planning and acquiring service
network capacity, and operating the network to fulfill customer requests.
There is a considerable amount of supply chain logistics going on in many
service organizations, but usually this is dwarfed by service response
logistics. The applicability of MS/OR to supply chain logistics is established
in every basic text on MS/OR. Applicability to service response logistics
is less strongly emphasized, but it is plain that many applications are
possible; for example:
-
forecasting models can anticipate customer demands for service;
-
inventory models can be adapted to service network capacity management,
where too little capacity results in lost service opportunities and too
much results in excessive costs;
-
location models can optimize the placement of service facilities;
-
optimization models can produce cost/service tradeoff curves and allocate
resources to facilities, equipment, and other components of the service-providing
infrastructure;
-
queueing models can facilitate managing waiting times at service facilities;
-
simulation and other models are useful for workflow re-engineering;
-
work force planning and scheduling models can improve personnel management.
Not only is MS/OR applicable, but the service sector seems to be receptive
to MS/OR techniques. The CLM (1991b) survey of banks, high-tech field service
organizations, hospitals, and telephone companies revealed that about 2/3
of those who had an opinion believe that mathematical and statistical models
and decision-support and expert-system applications will become more important
and useful in the next few years. The number who thought they would be
less so was negligible. Fully 72% of the respondents ventured an
opinion in the case of mathematical and statistical models, but only 32%
did so in the case of decision support and expert systems, a disparity
that may be turned to the advantage of MS/OR.
Thus, it appears that there are ample potential rewards awaiting the MS/OR
community if it were to focus a much larger share of its attention on the
service sector.
Shifting the MS/OR Focus. What could be done to bring about such
a shift of focus?
A good part of the answer lies with the universities, since research and
teaching seem to be lagging far more than actual practice. Curricula and
teaching materials need to be revised to reflect the ascendancy of the
service sector. R.P.I.'s new Center for Services Sector Research and Education,
and Wharton's Fishman-Davidson Center for the Study of the Service Sector,
apparently are the only two organized centers in the U.S. that focus on
service and have a significant MS/OR emphasis.
Another part of the answer lies with the agencies and foundations that
sponsor research. By comparison with manufacturing, relatively little support
is available for service-oriented research. A few funding initiatives in
this area could do much to capture academia's attention.
ORSA and TIMS also could do much to promote the reorientation of MS/OR
in the direction of service:
-
They could develop curriculum guidelines for achieving a more balanced
emphasis between the production and service sectors. To accelerate the
pace of change, this could be followed up with commissions or competitions
for service-oriented teaching materials.
-
They could sponsor meetings and publications that focus on service.
A newsletter, or even a dedicated section in OR/MS Today, would
be a good start.
-
They could give special nurturing to service-related ORSA technical
sections and special interest groups and TIMS colleges, forming new ones
where good opportunities exist.
-
They could lobby research-sponsoring agencies and foundations for new
programs and initiatives focusing on service.
-
They could commission a study to follow up on CLM (1991b) from the
MS/OR viewpoint. The aim would be to assess the potential role of MS/OR
in capturing the huge benefits claimed in that report for the concerted
application of logistics skills and techniques to the service sector, and
to prescribe specific steps by which this role could be realized.
-
They could seek to cooperate directly with the Council of Logistics
Management, the 6,000-member professional society that sponsored CLM (1991b),
on joint actions that would strengthen the service sector role of logistics
as a functional area and of MS/OR as a technical approach.
-
They could make the service sector a specific target for TIMS' public
relations effort in the future. Trade and professional associations that
cater to the service sector could assist in spreading the word if we can
convince them of MS/OR's potential contributions.
Optimal allocation of resources is supposed to be a specialty of MS/OR.
Let us not be guilty of grossly misallocating our academic and professional-society
resources.
Stress Embedded Applications
An embedded MS/OR application is a model and its solver, or a derivative
thereof, installed as an integral part of a functional host of some sort--such
as a business process, an information systems application program, a scheduling
procedure, process control software, or even a homogeneous group of front
line workers or a controller for a machine, appliance, instrument, or other
device. The application may execute model-based computations automatically
with the most current data whenever its host performs its function. Alternatively,
an embedded application may permit a considerable degree of discretion
as to how it is used (as distinct from whether it is used),
especially if there are many intended users.
The antithesis of an embedded application is the one-shot project, such
as a decision analysis of a regulatory issue or a simulation of a proposed
factory automation project. Whereas each one-shot project must be sold
individually, an embedded application need not be, in the sense of MS/OR
technology uses per sale. Once in place, it runs routinely and is no more
vulnerable than its host function is to budget cuts, the departure of a
management champion, organizational flattening, or staff headcount reductions
(the scourges of organized MS/OR groups for the last quarter century).
Moreover, embedded applications tend to keep MS/OR people in touch with
the host process through periodic maintenance and enhancement activities.
Examples. The use of linear programming for gasoline blending is
one of the oldest and best known examples of an embedded MS/OR application
(e.g., DeWitt et al. 1989). Other examples are: capital program management
at GTE (Bradley et al. 1986), crew scheduling and yield management at American
Airlines (Anbil et al. 1991, Smith, Leimkuhler, and Darrow 1991), production
planning and control of slab casters at Bethlehem Steel (Vasko et al. 1989),
real-time vehicle dispatch at Mobil Oil (Brown et al. 1987), and thermal
and hydro-unit commitment at SCS (Erwin et al. 1991).
John Little gives an appealing vision of embedded applications that empower
particular classes of workers. He sees important opportunities of this
sort emerging
... from the idea of empowering the organizational front
line. Classically we have built systems for end users largely as extensions
of ourselves; in other words, we solve the problem and deliver the solution
to others. A different goal is to give people systems with which they can
create their own solutions to problems. We can boost industrial productivity
by developing such tools for the salespeople, the telephone operators,
and the factory workers. In particular, if we can give them the continuous
means to monitor, understand, and improve their own performance, real gains
will follow. (Little, p. 541)
Among the examples he offers is "fact-based selling" systems for a sales
force. Notebook computers, the mass deployment of which began in 1991,
provide a good platform for such applications.
Little's vision can be viewed as the decision support system (DSS) concept
applied on a massive scale. Most existing DSSs would not be thought of
as an embedded MS/OR application, even when they contain models and solvers,
because they need not be used on a regular basis. But it is appropriate
to think of a DSS as embedded when it is used by all or nearly all of some
category of front line workers.
Opportunities for embedded applications have never been better, thanks
in substantial part to the computer/communications revolution; once a business
function is sufficiently computerized, it becomes ripe for improvement
via embedded decision technology. For proof, one need look no farther than
American Airlines Decision Technologies, which has built its spectacular
five-fold growth during the last three years largely on embedded applications
in the service sector (Cook 1991).
A fairly new opportunity for embedded applications arises from the proliferation
of microchips and microprocessors in all manner of devices. A familiar
example is provided by automobiles, which have had embedded microprocessors
for engine control since 1977 to optimize spark firing, ignition timing,
and fuel-injection (Minchillo 1991). Today's Nissan Infiniti Q45 has 14
separate computer modules controlling the adjustable suspension, antilock
braking, engine, power steering, security system, transmission, and more.
The Cadillac Allante has a true local area network linking most of its
22 microprocessors in place of the usual thick wiring harnesses. In a few
years, GM cars will contain a 32-bit chip based on the Motorola 68020,
which is the heart of a Macintosh II. With the equivalent of a Mac II,
a local area network, and at least one transceiver in the vehicle of the
future, opportunities for embedded mobile MS/OR technology abound (Spreitzer
1989).
Hardware Applications. One usually thinks of an embedded MS/OR application
as being embodied in software. But it can also be embodied in hardware,
such as a microchip that implements decision logic. This is becoming increasingly
practical as small volume integrated circuit design and fabrication technology
improves, and as microchips keep finding new uses in appliances, instruments,
machines, and other devices. I believe that great potential now exists
to put MS/OR technology into such objects to enhance their usefulness.
This has already begun to occur. For example, a heuristic for finding good
hole-drilling sequences for printed circuit board manufacture has been
put on a microchip and sold commercially in programmable drilling machine
controllers (Larson 1988).
The most spectacular examples are provided by various products introduced
in Japan, starting early in 1990. There is already a billion dollar market
there for consumer and industrial products incorporating fuzzy logic chips,
and one forecast has the market going to $15 billion by 1995 (JT 1991).
Matsushita, Mitsubishi, Sanyo, Sharp, Sony, and others have brought out
new products in dozens of different categories, and they are selling extremely
well in Japan.
To mention just two of these products: you can buy a fuzzy logic air conditioner
with 24% better energy efficiency, and a fuzzy logic elevator that reduces
average waiting times by 15-20% and greater-than-one-minute waiting times
by 30-40% (Armstrong and Gross 1990). There are chip makers specializing
in fuzzy logic chips, more than 2,000 patents have been issued in Japan,
and the Japanese government is subsidizing a Laboratory for International
Fuzzy Engineering Research at a level of $70 million for five years (Armstrong
and Gross 1990, Williams 1991).
I mention these remarkable developments not to promote more work in fuzzy
logic by our community, but rather to indicate the magnitude of the impact
that could possibly be made by embedding MS/OR applications into consumer
and industrial products. The Japanese boom in fuzzy logic products is based
largely on fuzzy if-then rule sets with no explicit analytical model in
the traditional MS/OR sense. The next step might well be to use model-based
decision technologies to improve product performance even more, and to
provide entirely new functionalities. If such a step could be taken
successfully, it would be difficult to imagine any other development that
could make a greater contribution to MS/OR's visibility or to its impact
on the national economy.
It has been argued that computers will truly have reached their zenith
when they have been embedded so subtly and so ubiquitously in everyday
objects that most people will cease to be aware of them; they will be aware
only of their protean functionalities as embodied by the computerized objects
(Weiser 1991). Similar arguments could be made that MS/OR applications
will have reached their zenith when they have been embedded so deeply in
everyday processes and objects that most managers and employees cease to
be aware of them even though they may encounter them often each day; people
will be aware only of their functionalities as embodied by the processes
they facilitate and the objects they vitalize.
Risks and Benefits. Embedded applications present important opportunities
for MS/OR, but they also involve some risks. One that has been recognized
for many years is that MS/OR can be embedded so successfully that MS/OR's
role in achieving the success will be soon forgotten. Hoffman (1975) identified
this as a difficulty with respect to middle management operational decision
processes at Standard Oil. He saw this occur with statistical forecasting
models, inventory control models, and many kinds of LP applications, with
IS taking these over. PoKempner (1977) also recognized this risk, calling
it an "into-the-woodwork syndrome" that can make MS/OR a victim of its
own success.
Thus, an important part of doing embedded applications successfully is
to take steps aimed at retaining indefinitely an explicit connection to
MS/OR. This seems like the professionally responsible thing to do anyway,
except when there is another technically qualified group well suited to
long term oversight. Then it may be appropriate to spin off the application.
Note that this is one of the ways in which MS/OR dispersion can occur.
I hope that the reader will not mistake my enthusiasm for embedded applications
as implying any disapproval for other kinds of applications. My main intent
is to make the point that there is a high-leverage, if somewhat less heroic,
way to succeed with MS/OR in addition to the stereotypical one that most
of us lust after: doing a large project with high-level sponsorship that
leads to compelling recommendations with huge verifiable savings.
Go Into the Quality Business
The quality movement is still very much on the upswing in U.S. industry,
after its storied successes in Japan. Originally centered in manufacturing,
service organizations have bought into quality in a big way: in a recent
survey of more than 200 service organizations, 2/3 reported that they have
formal quality programs (CLM 1991b). Informal discussions confirm that
interest in quality is growing rapidly in the service sector, a trend that
helps to connect this opportunity to one discussed previously.
The quality movement heavily emphasizes cultural changes for managers and
workers, but it still has a strong statistical orientation. For example,
there are elements of statistical quality control in about 80% of the examination
items for the coveted Malcolm Baldridge Award (Valenzuela 1990). Drucker
(1990) argues that statistical quality control is one of the four keys
to a successful manufacturing business.
MS/OR Opportunities. Since the quality movement is so vigorous and
makes such extensive use of statistics, it is only natural that MS/OR practitioners
should consider getting involved. Some have, with handsome success at some
companies (e.g., Valenzuela 1990, Albers 1991, Lucas 1991, Schwartz 1991).
Others have tried to move into quality without success.
Two observations are pertinent to the issue of MS/OR moving into quality.
First, conventional approaches to quality lead most employers to strongly
prefer statistical over MS/OR training. Unfortunately, statistical techniques
have received decreasing emphasis in many MS/OR educational programs over
the last couple of decades. The time has come for the pendulum to swing
back.
Second, there is controversy within the MS/OR community about whether quality
should be viewed as a legitimate domain for MS/OR. Some believe that the
quality movement is too mundane in technical content, and contains too
large a component of cultural intervention and faddishness. Others believe
that modern MS/OR technology can be adapted usefully to this purpose, and
that MS/OR professionals are obliged to help their organizations even for
applications that might turn out to be technically routine or to require
cultural reform. The former view seems more commonly found in academia
than in industry, but I join the second camp. MS/OR professionals are well
equipped to build models and help design processes that focus on quality
as a key objective.
Those who are concerned about the apparent lack of technical challenges
may take heart from the Board on Mathematical Sciences report which, as
noted earlier, placed "statistical methods for quality and productivity"
on its honor role of 27 "recent research accomplishments and related opportunities"
(BMS 1990). It discusses four subareas: statistical process control, statistical
experimental design, reliability, and acceptance sampling. Additionally,
modeling the interplay of cost and quality can be technically challenging
as well as of great practical importance (Cook 1991).
Educational Opportunities. A prerequisite for MS/OR to participate
meaningfully in the quality movement is much greater attention to quality
topics in MS/OR curricula at all levels. The time for this has long been
ripe, but progress has been disappointingly slow.
Operations management courses and curricula have been adding material and
courses on quality in recent years. According to Carraway and Freeland,
30% of the 20 leading MBA programs surveyed now have an elective course
on quality, and 90% of all required operations management courses now cover
aspects of quality. However, elective courses on quality, most of which
were added to curricula fairly recently, are among the least popular electives
(might this be a reflection of faculty attitudes?). And there is no indication
that quality is being covered in MS/OR courses at the MBA level, an impression
confirmed by perusal of popular introductory MS/OR texts.
A more recent survey of teaching and research in quality within business
schools in general and operations management in particular reveals a situation
regarded by some as scandalous (Kaplan 1991). For example: at "top twenty"
schools, an average of only 12% of available class time in introductory
operations management courses is spent on quality; and there were no articles
at all on quality in the 1990 crop of 278 articles in Management Science,
Journal of Manufacturing and Operations Management, and Operations
Research. Kaplan concludes that
... business schools completely missed the quality revolution
in management. ... While businesses throughout this country and the rest
of the world have mobilized their entire organizations on quality, business
school professors are still wondering whether quality improvement creates
value for organizations, whether research on quality is legitimate for
them to do, and whether they will get recognition and be rewarded for performing
such research (Kaplan, pp 19-20).
Yet major progress is being made at a few schools (Roberts 1991).
Opportunities for the Professional Societies. Universities may have
been slow to recognize the quality revolution, but professional societies
need not be. ORSA and TIMS could be taking a proactive role. Unfortunately,
there has been little or no attention as yet to quality opportunities at
the council level of either society, but grass roots momentum may be building.
For example, there were 15 sessions on quality at the Fall 1991 ORSA/TIMS
meeting compared with 2 five years ago and zero ten years ago; the Management
Science Roundtable made quality a theme of several recent meetings, with
the general conclusion that important opportunities await MS/OR professionals
who choose to work in this area; and CPMS included several sessions on
quality at its 1991 Summer Conference, with a similar conclusion.
One of the best steps that ORSA and TIMS could make at this time would
be to offer their cooperation to the Leadership Steering Committee of The
Total Quality Forum. The mission of this top-level group is to take the
quality revolution to universities because "... academic institutions that
are slow to embrace TQM, at best, miss the opportunity to lead change and,
at worst, run the risk of becoming less relevant to the business world"
(Akers et al. 1991, p. 94).
General. The MS/OR community may have a lot of catching up to do
in order to be truly useful to management in its quest for quality, but
there is no reason why we need to think small. Our objective need not be
simply to go from dragging our feet to pulling our weight. Our objective
could be as ambitious as making new model-based approaches the basis
for a new phase of the quality revolution. Such approaches can advance
quality goals via models that inform quality measurement, that lay bare
the web of influences on quality, and that drive or constrain on quality
measures.
An MS/OR campaign in the quality area would be strongly synergistic with
efforts to exploit the previous four opportunities, all of which point
toward front line workers, who are at the very center of the quality revolution.
Specifically: microcomputers and telecommunications empower the front line,
dispersed practitioners are closer to the front line than organized MS/OR
groups are, the service sector is traditionally more oriented toward front
line workers than the goods-producing sector is, and embedded MS/OR applications
tend to be closer to the front line than non-embedded applications are.
This concludes my survey of some of MS/OR's great opportunities to flourish.
CONCLUSION
My aim in writing this paper was to identify what appeared to me to be
the most influential forces acting on MS/OR, the most important current
trends, and some highly promising opportunities for the profession to flourish
in the future in light of these realities. There are bound to be differences
of opinion about my choices of forces, trends, and opportunities. I hope
that others will see fit not only to comment on my choices but also to
develop other topics for the MS/OR community to consider.
Some of the topics that I considered but chose not to develop here at length
are:
Forces: the competitiveness crisis in U.S. industry, economic globalization,
transition to a service and information based economy.
Trends: the increasing popularity of MS/OR in the mathematical community,
increasing industrial interest in statistical applications, possibly increasing
use of the MS/OR approach by consultants, increasingly narrow Ph.D.-level
education in MS/OR, decreasingly rigorous MBA-level education in MS/OR.
Opportunities: assimilate AI more aggressively, cooperate more closely
with Information Systems, stress cross-functional and environmental and
advanced manufacturing and public sector applications, enhance organizational
flexibility and responsiveness (a competitive necessity created by economic
globalization and the computer/communications revolution), improve the
conduct of MS/OR practice (e.g., more attention to the modeling life cycle
and greater exploitation of the object-oriented paradigm), improve the
conduct of MS/OR research (e.g., more empiricism, more attention to qualitative
issues, and better integration with behavioral science), and target applications
in the former Soviet block.
Discussing an opportunity and taking advantage of it are, of course, entirely
different things. There is no denying that it will take a great deal of
energy to capture the full potential of any of the opportunities that I
have discussed. This energy must be supplied by universities, MS/OR-using
organizations, professional societies, and, ultimately, by individuals.
Aside from the mass media--which cannot be mobilized on command--only ORSA
and TIMS have the leverage necessary to catalyze a major energy shift in
the direction of the great opportunities. Would you like to see this occur?
Would you be willing to help?
If so, I urge you to write to an officer of the professional society of
your choice. There is much to be done.
APPENDIX
A Company President Explains the Dissolution of the Central
MS/OR Group
What follows is part of a letter from the head of one of the world's largest
companies just prior to the final dissolution of its central MS/OR group
in the mid 1980s. The CEO at that time happened to have a masters degree
from a top university with an emphasis that included management science,
but that did not prevent the dissolution from occurring. Contrary to the
positive tone of this letter, the head of the dissolved group saw its demise
as a blow to MS/OR as a profession.
Dear Professor Geoffrion:
...Management science has made valuable contributions to
our business, and we continue to use it extensively throughout the company.
As [company name omitted] has grown and diversified, we
have become a decentralized corporation, having [number omitted] major
business units. Management Science has followed our organizational alignment
and has become integrated into the business units. There has, consequently,
been a steady migration of personnel from our corporate center of expertise.
This transition reflects our experience of how to implement management
science techniques most effectively, and we are now in the last stages
of implementation.
We think very highly of [corporate group leader name omitted],
and are in the process of identifying a suitable placement for him. The
mantle of corporate support for management science will fall upon our R&D
organization....
Yours very truly,
[name omitted]
President
ACKNOWLEDGMENTS
I am grateful to Omega Rho for its invitation to address the November 1991
ORSA/TIMS Meeting, which provided the impetus for preparing this paper.
I appreciate the help of the dozens of friends and colleagues who graciously
supplied criticisms of draft materials, examples, suggestions, and encouragement
during the gestation of this paper. Particular thanks are due to Hugh Miser,
whose editorial assistance surpassed even my high expectations, and to
the members of the Management Science Roundtable, who taught me most of
what I know about how practitioners view the profession and what is happening
in the world of practice.
REFERENCES
Akers, J.F., P.A. Allaire, E.L. Artzt, R.W. Galvin, H.A. Poling, and J.D.
Robinson III (1991). An Open Letter: TQM on the Campus. Harvard Bus.
Rev., 69(6), 94-95.
Albers, W.A. (1991). Getting Quality 'Under the Skin.' Presentation MC17.3,
ORSA/TIMS Joint National Meeting, Anaheim, November. (Head, Operating Sciences
Dept., General Motors Research Laboratories)
Anbil, R., E. Gelman, B. Patty, and R. Tanga (1991). Recent Advances in
Crew-Pairing Optimization at American Airlines. Interfaces 21(1),
62-74.
Armstrong L., and N. Gross (1990). Why 'Fuzzy Logic' Beats Black-or-White
Thinking. Business Week, Science & Technology Section, 21 May,
92-93.
Arvesen, J. (1991). Database Marketing in the Pharmaceutical Industry.
Presentation to the Management Science Roundtable, Tampa, February. (Director,
Department of Marketing Research and Planning, Pfizer Pharmaceuticals)
Bajgier, S.M., H.D. Maragah, M.S. Saccucci, A. Verzilli, and V.R. Prybutok
(1991). Introducing Students to Community Operations Research by Using
a City Neighborhood as a Living Laboratory. Opns. Res. 39(5),
701-709.
Banks, J., E. Aviles, J.R. McLaughlin, and R.C. Yuan (1991). The Simulator:
New Member of the Simulation Family. Interfaces 21(2), 76-86.
BLS (1990). Outlook 2000 and unpublished supporting tables. Bureau
of Labor Statistics, U.S. Government Printing Office, Washington, D.C.
BMS (1990). Renewing U.S. Mathematics: A Plan for the 1990s. Board
on Mathematical Sciences, National Academy Press, Washington, D.C.
Bradbard, D.A., F.N. Ford, J.F. Cox, and W.N. Ledbetter (1987). The Management
Science/Operations Research Industrial-Academic Interface. Interfaces
17(2), 39-48.
Bradley, G.H., T.M. Edwards, A.M. Geoffrion, R.D. Glomski, and T.R. Sweet
(1986). Optimizing the Capital Investment Portfolio. Proc. National Communications
Forum 40(1), 1-18.
Brown, G.G., C.L. Ellis, G.W. Graves, and D. Ronen (1987). Real-Time, Wide-Area
Dispatch of Mobil Tank Trucks. Interfaces 17(1), 107-120.
Buckholz, D.E. (1990). Communications, an Essential Requirement. Presentation
to the Management Science Roundtable, Phoenix, February. (Director, Operations
Research, Hallmark Cards, Inc.)
Bush, G. (1991). Economic Report of the President, U.S. Government
Printing Office, Washington, D.C.
Carraway, R.L., and J.R. Freeland (1989). MBA Training in Operations Management
and Quantitative Methods. Interfaces 19(4), 75-88.
CLM (1991a). Logistics Software, 1991 Edition. Prepared by Andersen
Consulting for the Council of Logistics Management, 2803 Butterfield Road
#380, Oak Brook, IL 60521.
CLM (1991b). Logistics in Service Industries. Prepared by Arthur
D. Little and The Pennsylvania State University for the Council of Logistics
Management, 2803 Butterfield Road #380, Oak Brook, IL 60521.
Cohen, M., P.V. Kamesam, P. Kleindorfer, H. Lee, and A. Tekerian (1990).
Optimizer: IBM's Multi-Echelon Inventory System for Managing Service Logistics.
Interfaces 20(1), 65-82.
CONDOR (1988). Operations Research: The Next Decade. Opns. Res.
36(4), 619-637.
Cook, T. (1990). Quoted in OR/MS: Looking Back on the 80s; Looking Forward
to the 90s. OR/MS Today 17(1), 26-29.
Cook, T. (1991). Personal communication, 11 October 1991. (President, American
Airlines Decision Technologies)
Davidson, J.D. (1981). Four Vice Presidents, Four Directors, and a Partridge
in a Pear Tree. Interfaces 11(1), 59-61.
Dertouzos, M.L. (1991). Communications, Computers and Networks. Scientific
American 265(3), 62-69.
DeWitt, C.W., L.S. Lasdon, A.D. Waren, D.A. Brenner, and S.A. Melhem (1989).
OMEGA: An Improved Gasoline Blending System for Texaco. Interfaces
19(1), 85-101.
Drucker, P.F. (1990). The Emerging Theory of Manufacturing. Harvard
Bus. Rev. 68(3), 94-102.
EDI Forum (1991). Special Edition: Introduction to Electronic Data Interchange.
EDI Forum, 4.
Efron, B., and R. Tibshirani (1991). Statistical Data Analysis in the Computer
Age. Science 253 (26 July 1991), 390-395.
Erwin, S.R., J.S. Griffith, J.T. Wood, K.D. Le, J.T. Day, and C.K. Yin
(1991). Using Optimization Software to Lower Overall Electric Production
Costs for Southern Company. Interfaces 21(1), 27-41.
Fylstra, D. (1991). The Democratization of Operations Research. OR/MS
Today 18(4), 12-13.
Gass, S. (1990). Model World: Danger, Beware the User as Modeler. Interfaces
20(3), 60-64.
Gray, P. (1979). Is OR/MS Everywhere?. Interfaces 9(5), 129-134.
Gries, D., and D. Marsh (1992). The 1988-89 Taulbee Survey Report. Commun.
ACM, 35, 133-143.
Hammer, M. (1990). Reengineering Work: Don't Automate, Obliterate. Harvard
Bus. Rev. 68(4), 104-112.
Hammond, J. (1974). The Roles of the Manager and Management Scientist in
Successful Implementation. Sloan Management Review (Winter), 1-24.
Harker, P.T. (1991). Services and American Competitiveness. Fishman-Davidson
Center for the Study of the Service Sector, The Wharton School, University
of Pennsylvania, July.
Hennessy, J.L., and N.P. Jouppi (1991). Computer Technology and Architecture:
An Evolving Interaction. Computer 24(9), 18-29.
Hoffman, G.M. (1975). The World of the Industrial Practitioner. Notes prepared
for a Panel Discussion, TIMS XXII, Kyoto, Japan, July.
Hoffman, G.M. (1991). What Industry Wants From the Universities. Working
Paper 91/13, McCormick School of Engineering and Applied Sciences, Northwestern
University, 21 August.
Hopper, M.D. (1990). Rattling SABRE--New Ways to Compete on Information.
Harvard Bus. Rev. 68(3), 118-125.
Jarvenpaa, S.L., B. Ives, and G.B. Davis (1991). Supply/Demand of IS Doctorates
in the 1990s. Commun. ACM 34(1), 86-99.
JT (1991). Many Appliance Makers Are Vibrating to the Latest Buzzword;
EURO Fad May Follow the Fuzzy Boom. Japan Times Weekly International
Edition, Focus, January 21-27.
Kaplan, R.S. (1991). The Topic of Quality in Business School Education
and Research. Selections, Graduate Management Admission Council,
Autumn, 13-21.
Keen, P.G.W. (1988). Competing in Time: Using Telecommunications for
Competitive Advantage. International Center for Information Technologies,
Washington, D.C.
Keen, P.G.W. (1991). Personal communication, 7 November 1991. (Executive
Director, International Center for Information Technologies)
Kirkwood, C.W. (1990). Does Operations Research Address Strategy? Opns.
Res. 38(5), 747-751.
Larson, R.C. (1988). Operations Research in the Service Industries. In
Managing Innovation: Cases From Service Industries, B.R. Guiles
and J.B. Quinn (eds.). National Academy Press, Washington, D.C., 115-143.
Larson, R.C. (1990). The Queue Inference Engine: Deducing Queue Statistics
From Transactional Data. Management Science 36(5), 586-601.
Lilien, G.L. (1987). MS/OR: A Mid-Life Crisis. Interfaces 17(2),
35-38.
Little, J.D.C. (1991). Operations Research in Industry: New Opportunities
in a Changing World. Opns. Res. 39(4), 531-542.
Llewellyn, J., and R. Sharda (1990). Linear Programming Software for Personal
Computers: 1990 Survey. OR/MS Today 17(5), 35-47.
Lucas, J.M. (1991). Customer Service in Consulting. Presentation to the
Management Science Roundtable, Anaheim, California, October. (Technical
& Professional Leader, Consulting Division, Electronic Data Systems
Corp.)
Machamer, A., and R. Smith (1980). Career Planning and Management Development:
The Unrecognized Potential of Centralized OR/MS. Interfaces 10(3),
71-75.
Malone, T.W., and J.F. Rockart (1991). Computers, Networks and the Corporation.
Scientific American 265(3), 128-136.
Malone, T.W., J. Yates, and R.I. Benjamin (1989). The Logic of Electronic
Markets. Harvard Bus. Rev. 67(3), 166-172.
McElroy, M.P., and L.L. Trafton (1991). Network Modeling of DEC's Manufacturing
and Logistics Material Flows. Presentation at the Annual Conference of
the Council of Logistics Management, New Orleans, October.
Minchillo, V. (1991). Chips in the Old Block, sidebar of R. Cook, "Embedded
Systems in Control." BYTE 16(6), 153-160.
Mintzberg, H. (1973). The Nature of Managerial Work. Prentice-Hall,
Englewood Cliffs, NJ.
Miser, H.J. (1976). Introducing Operational Research. Opnal. Res. Quart.
27(3), 655-670.
Morgan, C.L. (1989). A Survey of MS/OR Surveys. Interfaces 19(6),
95-103.
MSR (1986). Problems and Opportunities Facing MS/OR. Special Report to
TIMS Council, Management Science Roundtable, March.
Nigam, R. (1989). Management Science in Financial Services. Presentation
to the Management Science Roundtable, New York City, October. (Chief Scientist,
Merrill Lynch & Company, Inc.)
NRC (1989). Everybody Counts: A Report to the Nation on the Future of
Mathematics. National Research Council, National Academy Press, Washington,
D.C.
OOQ (1990). Occupational Outlook Quarterly 34(1) (Spring),
Bureau of Labor Statistics, U.S. Government Printing Office, Washington,
D.C.
ORSA (1991). Final Report of the Operations Research Review Committee.
Submitted to ORSA President C.M. Harris by D.Y. Burman, H. Emmons, A. Geoffrion,
J.A. Muckstadt, S.M. Pollock, and R.E. Rosenthal, May 2.
Pister, K., et al. (1991). Report of the Universitywide Task Force on Faculty
Rewards. Office of the President, University of California, Berkeley, June
26.
PoKempner, S.J. (1977). Management Science in Business. The Conference
Board, Inc., New York.
Pool, R. (1990). Who Will Do Science in the 1990s? Science 248
(April 27), 433-435.
Quinn, P., B. Andrews, and H. Parsons (1991). Allocating Telecommunications
Resources at L.L. Bean, Inc. Interfaces 21(1), 75-91.
Radnor, M., and R.D. Neal (1973). The Progress of Management Science Activities
in Large U.S. Industrial Corporations. Opns. Res. 21(2),
427-450.
Ranyard, J.C. (1989). The Effective Practice of OR in the UK. Presentation
to the Management Science Roundtable, New York City, October. (Manager,
OR Executive, British Coal; President, Operational Research Society)
Rash, W. (1991). Corporate Connections. BYTE 16(9), 215-223.
Reich, R.B. (1991). The Work of Nations. Knopf, New York.
Roach, S.S. (1991). Services Under Siege--The Restructuring Imperative.
Harvard Bus. Rev. 69(5), 82-91.
Roberts, H.V. (1991). Practicing TQ To Learn TQ. Paper presented to the
Procter & Gamble Total Quality Forum, Cincinnati, Ohio, August 7.
Ryan, R. (1991). Downsizing: Bane or Boon? BYTE 16(9), 227-232.
Samuelson, D. (1990). Labor Dept. Study: Major Changes in OR Opportunities.
OR/MS Today 17(4), 22.
Schwartz, L. (1991). Customer Service in MS Groups. Presentation to the
Management Science Roundtable, Anaheim, California, November. (Director,
Strategic Information Technology, Pitney Bowes)
Science (1991). Science Careers. Special pull-out section, Science 252
(May 24), 1110-1148.
Smith, B.C., J.F. Leimkuhler, and R.M. Darrow (1991). Yield Management
at American Airlines. American Airlines Decision Technologies, Dallas/Fort
Worth Airport.
Spencer, T., A.J. Brigandi, D.R. Dargon, and M.J. Sheehan (1990). AT&T's
Telemarketing Site Selection System Offers Customer Support. Interfaces
20(1), 83-96.
Spreitzer, W.M. (1989). Technology, Vehicles, Highways and Future Transportation.
GMR-6878, General Motors Research Laboratories, Warren, Mich.
Swain, J.J. (1991). Simulation Survey: World of Choices. OR/MS Today
18(5), 81-102.
Teach, R. (1991). State of the Profession. OR/MS Today 18(5),
26-30.
Valenzuela, C.A. (1990). Implementing Quality Improvement Programs: A Challenge
for Management Science. Presentation to the Management Science Roundtable,
Philadelphia, October. (Manager, Applied Mathematics and Statistics, Air
Products and Chemicals, Inc.)
Vasko, F.J., F.E. Wolf, K.L. Stott, and J.W. Scheirer (1989). Selecting
Optimal Ingot Sizes for Bethlehem Steel. Interfaces 19(1),
68-84.
Von Simson, E.M. (1990). The 'Centrally Decentralized' IS Organization.
Harvard Bus. Rev. 68(4), 158-162.
Wagner, H.M. (1988). Operations Research: A Global Language for Business
Strategy. Opns. Res. 36(5), 797-803.
Weiser, M. (1991). The Computer in the 21st Century. Scientific American
265(3) (September), 94-104.
White, J. (1991). An Existence Theorem for OR/MS. Opns. Res. 39(2),
183-193.
Williams, T. (1991). Fuzzy Logic Simplifies Complex Control Problems. Computer
Design, March, 90-102.