Stephen M. Pollock is Herrick Emeritus Professor of Manufacturing and Professor Emeritus of Industrial and Operations Engineering at the University of Michigan. He has been involved in applying operations research and decision analysis methods to understand and influence of a variety of operational phenomena, including military search and detection, criminal recidivism, manufacturing process monitoring, sequential allocation of resources, predictive and proactive maintenance, networks of queues, the stochastic behavior of infectious disease epidemics, and the optimization of radiation oncology plans.
After receiving his Ph.D. at M.I.T. in 1964 he was a member of the technical staff at Arthur D. Little, Inc. before joining the faculty at the U.S. Naval Postgraduate School in 1965 and the University of Michigan in 1969. He was chair of the IOE Department from 1981 through 1990. He has chaired the University's Research Policies Committee and Tenure Committee, and was a member the College of Engineering's Executive Committee. In 1992 he was the recipient of the Stephen S. Attwood Award, the highest honor awarded to a faculty member by the College of Engineering. He has authored over 60 technical papers, co-edited two books, and has served as a consultant to over 30 industrial, governmental and service organizations.
Professor Pollock has been Associate Editor and Area Editor of Operations Research, Senior Editor of IIE Transactions, Associate Editor of Management Science and on the editorial boards of other journals. He has served on various advisory boards for the National Science Foundation, the Army Science Board. He has served on (and chaired) various National Research Council panels, and was a member of its Committee on Applied and Theoretical Statistics. He was President of the Operations Research Society of America in 1986 and awarded the 2001 INFORMS Kimball Medal for contributions to operations research and the management sciences. He a fellow of INFORMS and the AAAS and is a member of the National Academy of Engineering.
STEVE POLLOCK AND MICHIGAN
When I was very young I wanted to be a “scientist”, which I imagined meant having a laboratory in a small building behind my house, where I would do “experiments”. Of course, I had no idea what that meant – perhaps mixing acids and alkalis, maybe melting stuff or something to do with gears or electricity. Then around the sixth grade I was introduced to “word problems” in math class: “Pat rows at 25 feet per minute across a 200 foot wide river that flows at 5 miles per hour. How far downstream …?” Most of my schoolmates were terrorized by these primitive exercises in mathematical modeling, but I was perversely fascinated. Not by the math itself, but by the idea that you could actually use math to solve a problem of possible interest – to rowers, perhaps. Although my approaches to these early examples of mathematical models may have been weird, they were usually not completely wrong.
The vagaries and realities of academic counseling in the Brooklyn N.Y. public school system led to my becoming a student at Cornell, majoring not in applied math but in a newly created field: Engineering Physics. Transistors had been just invented and this new discipline was aimed at producing the next generation of solid-state physicists. As it turned out, most of the students in this small program did not end up contributing to the semiconductor revolution, although it did produce an early generation of operations research (“OR”) analysts, including Al Blumstein and Bill Jewell … and me. But the experience did little to reduce my naivety: many years after I graduated I discovered that Professor Bethe, my undergrad counselor at Cornell -- a kind, attentive fatherly professor with an accent who signed my course election forms every term – was the Hans Bethe.
Being unaware that my undergraduate education was not really about physics but was more about analytical approaches to problem solving, ignorance and academic inertia led me to pursue graduate studies in physics at MIT. Although, as a physics grad student I greatly enjoyed climbing around MIT’s cyclotron and Brookhaven’s Bevatron, it was clear to me (and much clearer to my professors) that I was not cut out to be a physicist. As an experimentalist, after gathering data from hundreds of thousands of spark chamber images, I made the mistake of questioning why one used “least sum of square deviations” to fit an empirical curve rather than, say, using “least sum of absolute value deviations”. And the theoretical physics professors did not appreciate my dismissing the Schrödinger’s cat paradox as being simply a misunderstanding of subjective probability. Moreover, I severely struggled through my physics coursework and examinations. Fortunately, through the intervention of Philip M. Morse, an eminent physicist (and at the time Chair of the MIT Physics department) and one of the founders of the Operations Research Society of America, I made use of a policy (one that also exists at Michigan) that allowed me, after eventually passing the physics department’s doctoral exams, to be free to find a dissertation topic as long as I could put together a committee. I did, and the topic was -- a word problem (!) dealing with probabilistic search. Ron Howard (the co-director of MIT’s newly formed Operations Research Center, and soon-to-be founder of the OR sub-discipline of Decision Analysis) agreed to become my research advisor. Phil Morse chaired my committee. (Another member was Herbert Galliher, who left MIT in 1963 to join Michigan’s IE department.)
Shortly before I got my degree, my mentors suggested I pursue a career in which I could actually be paid for solving word problems. They suggested that I join Arthur D. Little, at the time one of the few non-governmental organizations having a large group dedicated to OR (and the newly named variant Management Science). The job was fascinating and varied. Clients of projects I worked on included Johnson and Johnson (inventory control), Anheuser-Busch (production scheduling), the US Navy’s Trident submarine program (anti-submarine warfare) and the US Air Force (evaluating primitive Russian-to-English computer translation software). Guidance at ADL was provided by early OR luminaries George Kimball, Martin Ernst and Bernard Koopman. However, after a few years of pursuing and landing clients, and lots of travel, a switch to academia became attractive and more controllable than the life of a consultant. So in1965 I jumped at an opportunity to become part of the newly formed Department of Operations Analysis at the US Naval Postgraduate School, in Monterey CA.
Coming to Michigan
The four years spent in Monterey were, to say the least, “interesting”. These were the hectic flower-power times of the late 1960’s, and my students were military officers, most of whom were either coming back from or going to Vietnam. Although I had a big peace symbol on my car, I also strongly felt that -- since our country will always have a military -- my contribution would be to help ensure that the decision making of our young officers would be grounded in reality and as effective, as possible (I still think so). On the other hand, after three years I was inspired by a talk by Al Blumstein, who was on a mission to show that OR methods could be used not only to address and analyze problems in the (at that time) “traditional” arenas of the military, production and commerce, but also to attack newly identified challenges in the public sector: improving police and fire operations, water distribution, health care, environmental protection, public transport, health systems, sports, etc.
Perhaps sensing my ambivalence, Seth Bonder contacted me in early 1969. Even though his own research group was heavily involved in military OR projects, he convinced me that if I joined the Department of Industrial Engineering (as it was named until 1972, when it was changed to Industrial and Operations Engineering – IOE) I would be able to work on whatever OR problems interested me, and in particular those involving the public sector. My interview talk was on methods of evaluating information retrieval systems. Although the talk must have been OK (I was hired) I never did work in that area again. But Herb Galliher, one of my MIT mentors, was doing seminal work on determining the best frequency, as a function of age, for breast cancer screening, and pap smears -- areas which are still open to research investigation; Clyde Johnson and Walt Hancock were developing the new discipline of hospital systems engineering; links were being forged with the Institute for Public Policy; novel optimization algorithms were being developed and expanded by Katta Murty and others; and Bonder’s military studies continued – eventually at his spin-off company. After a few years I discovered the perfect answer to the question commonly asked me: “Why in the world would you, in your mid-30’s, leave beautiful Monterey California to come to Michigan? The answer was clearly “I was too young to retire”.
The atmosphere at Michigan
In the many years from 1969 through the present, UM IOE has been my intellectual, professional and collegial home. The features of the department that had attracted me have been carried (and in most aspects, strengthened) through the first decade of the twenty first century. These unique attributes include:
• A collegial appreciation that academic research can be not only interesting, challenging and rewarding, it can also be fun.
• An understanding that the word “elite”, when associated with our university, is not a pejorative term. I am convinced that, over the past four decades, Michigan faculty (particularly those in the College of Engineering) have been fully aware the top 20% of our undergraduate students can match up – in intellectual ability, creativity, leadership qualities and social awareness -- with the very best students at any other educational institution in the country. For a public university this is rare, and demonstrated by our peer institutions every year by their trying to attract our best undergraduates to join their respective graduate programs. I have never been embarrassed to explain to prospective students, alumni, colleagues at other institutions and complete strangers that the University is indeed an “elite” institution.
• The success our department has had in attracting and accommodating faculty, students and research staff with a wide variety of interests: from the rather abstract mathematics of non-linear optimization and stochastic processes to the concrete experimentation involved in human factors or manufacturing systems analysis. That we have been ranked, by various sources, consistently among the top two or three among our peer programs, is something about which I am exceedingly proud to have been part of the process that led to such recognition.
• The ability to link up with other units and individuals across the University – whether for research or teaching – has been relatively barrier-free.
• Graduate students were not pre-selected to work with particular faculty or on particular projects; the student-advisor “marriage” process was instead informed by consensual agreement on matches of skills and interests, problem identification and resources, goals and timelines between student and faculty advisor. (Unfortunately, the exigencies of contemporary doctoral student-funding models have practically destroyed this intellectually valuable feature.)
• The presence of exemplary Deans of the CoE, particularly those during the 1980’s and 90’s. Jim Duderstadt was not only a master administrator; he was (and is) a visionary. I can still recall his striding down the hallway of the East Engineering building in the early 1980’s with a computer in his arms (was it a Compaq? A Commodore?), explaining to all he passed that by the end of the decade there will be these “portable” (his must have weighted 20 pounds) computers every classroom; and maybe eventually even in one’s home. Chuck Vest was an equally talented motivator, recognizer of faculty talent, and consensus-developer. It was a great pleasure to be able to work under them.
• There is probably no other institution in the world that would have allowed, supported and rewarded my research agenda. I was never discouraged or unsupported in my quest to explore a ridiculously wide variety of (word) problems. The resulting analyses, both theoretical and applied, arguably have led to a better understanding of a wide and disparate set of issues. These included analyses of:
Public Systems (broadly defined): Feedback in Social Systems;Energy R & D Modeling; Decision Analysis in China; Political Redistricting; Characterizing Adversaries of Nuclear Materials Safeguards Systems; Criminal Recidivism; Rural Police Patrol; Collusion in Bids for Metal Pipe and Road Salt; Sewage Plant Operations; Terrestrial Irradiation.
Medical/health: Diagnosis of Diabetic Patients; Staff Levels, Cost and Productivity of Hospital Units; Occupational Injuries; Early Infection in the Spread of HIV; Adaptive Radiotherapy Under Uncertainty; Muscle Fatigue in Cyclic Exertions.
“Traditional” Industrial Engineering: Coordinate Measurement and Circularity; Ship Pipe Routing; Queueing Networks with Blocking; Whether Components or Assemblies Should be Made to Specifications; Weighted Batch Means for Simulations; Testing of Motor Vehicle Emissions; Canning Problems; Classification of Spot Welds; Maintenance of Machines; Type Mating in Manufacturing.
Situations that just “came up”: Assignment of Swimmers to Events; The Optimality of Having Two Sexes; Golf Handicapping; Sequential Outlier Rejection.
• The general atmosphere that encouraged collaboration led to wonderful research adventures leading to publications with over close to a dozen IOE faculty colleagues, and more than 10 from other departments and units.
The Mathematical Modeling Studio
In the early 1970’s, Seth Bonder created “Concepts in Mathematical Modeling of Large Scale Systems”, a seminar type course in which students were asked to develop models of poorly specified but real “operational situations”. The goal was not to create “solutions” to problems, but to understand the process by which the mathematical representation of these operational situations could be formulated, simplified, tested and exposited, and – hopefully – used in practice and transferred to other practitioners. Seth and I collaborated on it for a few years, and in the early 1970’s, after he left to run Vector research Inc., I assumed responsibility for it (which I jealousy guarded for over three decades!). This course evolved into what is now called the “modeling studio”. Pride in this course (and its principles) prompts me to leave humility aside and brag that many ex-students who suffered through its demanding time and creativity requirements have written to say it was the most influential educational experience they had at Michigan. Equally satisfying is to know that colleagues throughout the country have adapted versions for use at their own institutions. I even manage to publish a few papers about it. Since my retirement professor Amy Cohn picked up the cause, which augurs well for the education of future operations engineers.
As my wife Tina has often observed over the past few years, I have clearly “flunked” retirement. Consulting opportunities allow me to help inform decision makers dealing with important issues. These, along with serving on various committees (particularly for the National Research Council) have not only kept me busy, involved and challenged, they also remind me that solving once daunting “word problems” can actually be an enjoyable retirement activity, even though there is no laboratory in my back yard. Maybe in my mid-70’s I’m still just “too young to retire”.
From my vantage point, the challenge to the University, the College of Engineering and the Department of Industrial and Operations Engineering of maintaining these strengths in the 20-teens is daunting. When students can instantaneously use the web to find “factual” answers to almost any question, our pedagogy must evolve to encourage the relevance of figuring out what questions to ask. The student-to-faculty ratio is increasing at the same time research resources per faculty member are decreasing. We need to rediscover a good balance between research load and teaching opportunities, between education and training, between specialization and flexibility, between following the latest research-fad-money direction and exploring interesting and potentially important issues. With our recent successful departmental administrative transition, a continued renewal of junior faculty and flow of excellent and motivated students, we are well positioned to adapt and excel in our educational mission to be the world’s best department of Industrial and Operations Engineering.