THE SPECIALIST WITH A UNIVERSAL MIND

ANDREW VAZSONYI, Feature Editor, McLaren School of Business, University of San Francisco

Operations management is just about the hottest subject right now. Goldratt's novel, The Goal, sold 800,000 copies. Zangwill published a paper on Japanese production in the September-October 1992 issue of Interfaces and the editor-in-chief, Fred Murphy, tells me he received more comments on the article than on any other article he ever published. The Wall Street Journal keeps on publishing the subject.

POM poses special problems to schools of management. Many schools are dropping management science and putting quantitative techniques into POM courses. This results in de-emphasizing decision sciences. Samaddar's article is an important contribution toward establishing our position with regard to teaching POM.

A Model of POM Curriculum: Rediscovering Taxonomical Balance

by Subhashish Samaddar, Western Illinois University

Let me first point out that there is an abundance of quantitative models for Production/Operations Management (decision making, at least that is what they are intended for). However, relevance of these models to POM should not be taken for granted. Influenced by their academic and professional background, and according to their perceptions of what a POManager ought to know, POM teachers incorporate a subset of these models in their models of POM. In recent years, however, there has been a notable variety in the latter. A POM course, on one extreme, could explore a very qualitative or behavioral type understanding of operations management; on the other extreme, it could dwell on all the quantitative models that exist under the sky (the latter ignores the existence of MS/OR courses in a business school). Recent textbooks in POM are living proofs of the dilemma of developing the correct mix of quantitative models and qualitative concepts.

A Taxonomical Model of POM

Figure 1 represents the model. At present, the content of a POM course takes the following form:

POM = a times A + b times B

(1)

where A represents the set of qualitative concepts (such as strategic planning, human and behavioral factors and so on) that are relevant to POM and a represents the proportion of the same picked by an instructor for his/her course; B and b represent the same for quantitative models; and a+b=100%.

The quantitative models (B) for POM can be further classified into three groups according to their genetic origination. Let the first group include models from MS/OR, viz. LP, IP, Simulation, Queuing Theory, PERT/Networks, Inventory; and I will call this group the Hard Group (X). The Medium Group (Y) includes models that come from Statistics and Econometrics, viz. Quality Control, Power Curves, Sampling Sizes, Regression, and Forecasting. Finally, the Soft Group (Z), if you will, includes Break-even-point, Make-or-buy, Cash-flow, NPV, Layout Design type models. So, according to this classification:

b times B = x times X + y times Y + z times Z,

(2)

where x, y, and z are, respectively, the proportions of X, Y, and Z that a POM teacher picks. A recently published longitudinal survey (Lane et al, 1993) reports,"Three quantitative techniques stand out as consistently believed to be the most important: math programming, statistics, and simulation. Other techniques vary in relative importance. Practitioners indicate the use of a more diverse set of techniques than educators."

Some Observations

A recent tendency among POM teachers is drastic reduction of b in favor of a in model (1). Call it a fad, or a taboo, models from above hard and medium groups (i.e., X and Y) are being shunned away. Following this mood POM textbooks in their recent versions are tucking the models from X and Y groups into appendices or supplements. Some books have even chosen to drop such models altogether.

This is not to imply that X and Y models do not need to pass the test of relevance to POM. Any model or its extension that fails such a test should become a good candidate for removal. For example, one may not see much use of steady state analysis with Poisson and exponential distribution in real life POM. Or consider the extension of PERT under uncertainty; the assumptions are unrealistic and it is very hard, if not impossible, to find bona-fide commercial applications of uncertainty formulas of PERT.

I believe in the academic freedom of an instructor to choose the content of his/her course. However, I also believe that it is high time that we debate the issue of the proper balance of content of a POM course, and try to converge, as a community, on a common subset of concepts and models. I leave that task to everyone interested in this field. Let me conclude this section by offering some anecdotal evidences to illustrate the above tendency.

Two years ago, on interviewing a candidate for a POM position, I inquired about his/her thoughts on the balance of a and b in model (1). Answer was, "Quantitative models (in POM) are a matter of history anymore ... ." A little bewildered, I asked for an example case. "For example, consider LP. Why should we teach LP in POM ... they can get that in MS/OR, and more importantly, we have excellent software that do all these `hard' stuff ... ." Recognizing that the candidate was rejecting the importance of a modeling approach because of its complex solution methods, I indicated that a POM course may not include methods such as Simplex. However, a two product (i.e., a two variable, and hence graphically solvable) resource constrained LP problem can be used nicely to give the students the idea of solution space (various mix of products), constraints (resource limitations), shadow prices (where to best spend the next dollar); and sometimes counter intuitive optimum results (for example, making an item with less profit contribution more in the mix realizes an overall maximum profit). Then, I referred the candidate to the POM text book by Krajewski and Ritzman that has a good treatment of LP in POM. (Reader: there are other books that do equal justice to this topic, I mentioned this book because I have used it in the past.)

Concluding Remarks

I know that quite a few of us think in the same vein as this POM teacher did. However, I believe we should do better than that. Instead of dropping quantitative models altogether from POM as a result of their algorithmic complexities, we need to bring out and demonstrate the novelty of applying these "hard" modeling approaches, the importance of the insights that these models, when formulated and solved, have to offer. Clearly, the onus is on the POM teacher--otherwise who will? It will take proper training, motivation to do justice to the field, hard work to design better presentations, and more importantly, the knowledge about the emphasis of the fieldþwhich, I believe, is making better decisions. Presentations can be helped by using, among others, transient solutions and prototyping. Furthermore, in addition to verbal lectures, the use of visual (such as colored graphics) and simple physical objects (such as Lego toys for expressing the idea of assembly and cycle time, and so on) can be very helpful.

Finally, not only should we focus on, and hopefully resolve the debate of quantitative and qualitative content of POM, we also need to think about another requirement of the field. We all agree that collection, management, and meaningful use of information play a critical role in managing anything, including Productions and Operations. Without relevant information and well managed data, decision-making models are like guns without bullets. This is more so in the 1990's due to the growing importance of integrated systems such as Computer Integrated Manufacturing/ Service that critically depend on managing tremendous amount of timely and complex data. Likewise, recent thoughts concerning combining POM and MIS research (Showalter, 1992) are timely and purposeful. Practitioners are also voicing the same need. For example, Frank Skidmore, the Director of Quality and Information Technology at IBM, and a Baldrige examiner himself, perceives information management as the "glue" that holds TQM systems together (Skidmore, 1993). Observing the importance and co-relevance of these two fields (i.e., POM and MIS), and drawing upon my own experience with teaching courses in both areas, I would like to propose a future model of POM:

POM = a times A + b times B + c times C

(3)

where C includes MIS topics such as basic designs and capabilities of information systems, available powers of hardware and software, information literacy versus data literacy, integrated systems, etc.

References

Krajewski, L. J. & Ritzman, L. P. Operations Management. Reading, Massachusetts: Addison-Wesley, 1993.

Lane, M. S., Mansour, A. H., & Harpell, J. L. Operations research techniques: A longitudinal update 1973-1988. Interfaces, 1993, 23(2).

Showalter, M. J. Integration of P/OM and MIS research. Decision Line, 1992, 23(4).

Skidmore, F. H. Information for TQM: Ready...Fire...Aim.... Proceedings of TQM: Pitfalls & Prescriptions, CIBER of Southern Illinois University at Carbondale, St. Louis, MO, 1993.

SUBHASHISH SAMADDAR is a faculty member in the Management Information Resources Department of the College of Business, Western Illinois University. He holds a terminal degree in operations management. Professor Samaddar's teaching and research interests include both POM and MIS.