October 1996 € Volume 23 € Number 5

Issue in Education:
Software Decisions

---

By Wayne L. Winston

The instructor of an introductory OR/MS course has a critical decision to make: Which software to use? To my mind, this decision depends on the type of class the instructor is teaching. Is it the required undergraduate or MBA management science course (taken primarily by non-quantitative majors)? Is it the undergraduate or master's level management science or OR course required for a quantitative major (in OR, industrial engineering or quantitative analysis)? Or is it an executive education MS course?


The required MS course for undergraduates and MBAs (non-quantitative majors)
At most schools, this course will be the student's only exposure to management science. I believe the only course that makes sense for these students is a spreadsheet-based application course with no discussion of algorithms. Most students in this course will have had at least five weeks of spreadsheet experience in their required computer course, so a spreadsheet-oriented course builds on and enhances prior spreadsheet skills. I would primarily emphasize optimization, simulation, queuing and decision making under uncertainty. Such a course has several advantages:

• It develops the student's spreadsheet and modeling skills. Ninety-five percent of students in this course will use a spreadsheet as their primary vehicle of analysis in the workforce, so anything we can do to make them see the power of the spreadsheet is a real plus.
• Such a course develops skills the student can use in their upper-level finance, marketing and operations courses. For example, queuing models can be used for back-of-the-envelope calculations in discussions of process redesign; and simulation can be used in marketing and finance classes to analyze the introduction of a new product and make capital budgeting decisions. By encouraging the use of modeling, we are the engine that drives an improved curriculum. This will certainly raise our (usually low) status on the business school "totem pole."
• Students think better in a spreadsheet than with the traditional "x, y and z" approach (see Winston [1996] for further discussion of this matter).
• With a spreadsheet we can show the student the power of OR/MS, ensuring that when they enter the workforce they will seek out OR/MS professionals to help them solve problems. After showing students how to price options with simulation, rate sports teams, or solve a facilities location problem, students will understand the power and utility of our discipline. We can be the engine that drives other disciplines to adopt the OR/MS paradigm for problem solving. Students should be cautioned, however, that many industrial-strength problems cannot be solved with spreadsheets.
• A spreadsheet enables us to integrate the data analysis and model analysis (simulation or optimization) in a single workbook. Certainly the spreadsheet is the only viable option for a course in which statistics and MS are both taught.

I just don't see the traditional algorithmic approach delivering the goods like the spreadsheet approach. How many marketing and finance majors will use the simplex algorithm in their jobs?


The required MS/OR course for quantitative majors
For students majoring in industrial engineering, operations research or quantitative methods, I believe the traditional algorithmic course (with a moderate dose of spreadsheets) is still the way to go. Who will write the computer codes and develop the improved algorithms for the future if we do not teach algorithms in these courses? Most students in these courses have (or should have) some kind of computer programming background.

As the primary software in such a course, I would use a modeling optimization package such as LINGO, AMPL or GAMS, and a simulation package such as SLAM and SIMAN. I would also introduce the students to spreadsheet solvers and a spreadsheet simulation add-in (@RISK or Crystal Ball). I think it is important to address interesting applications in this type of course to motivate the students to learn the math behind the algorithms.


Executive education MS courses
Most marketing and financial analysts with major corporations graduated before spreadsheets became popular. While these analysts use spreadsheets, many are uncomfortable with them. We have a huge opportunity to increase the "human capital" of these analysts by training them in data analysis and spreadsheet model-building.

I have taught such a course with great success at Eli Lilly. We concentrate equally on optimization and simulation applications. As a byproduct of the course, the analysts feel much more comfortable with building models that realistically incorporate the day-to-day uncertainty faced by their company.

Again, we must realize that for most corporate analysts, their sole tool for analysis is the spreadsheet. As one employee of McKinsey told me, "We live in spreadsheets." The sooner we realize this, the more successful we will be.

The next decade should be the time when non-OR/MS professionals learn how to unleash the power of mathematical modeling. The spreadsheet should be the primary vehicle used to accomplish this worthy goal.


Reference
1. Winston, W.L., "Management Science with Spreadsheets for MBAs at Indiana University," Interfaces, Volume 26, Number 2, 1996, pages 105-111.

Wayne L. Winston is a professor of Decision Sciences at Indiana University School of Business, Bloomington, Ind.