Kellogg logo

Decision Sciences D36

[JUMP TO THE CURRENT TOPIC]

This is the course page for Decision Sciences D36 (DS-D36), Accelerated Mathematical Methods for Management Decisions. It will be taught by Roger Myerson, at the Kellogg Graduate School of Management, Northwestern Unversity (Evanston campus), in Fall 1999.

Class meetings will be Mondays and Wednesdays 7:00-8:40 p.m. in Leverone G05.
Office hours: Tuesdays 12-4pm in Leverone 569.
Review sessions with Deepasriya Sampath Kumar will be Wednesdays, 9:00-10:30 a.m. in Leverone 101.

Test Schedule: A 1-hour in-class quiz on the first two homeworks is now scheduled for Monday October 18. A longer take-home midterm is scheduled for November 3-8, and the take-home final is scheduled for December 1-6.
The take home final should be returned to Roger Myerson in the MEDS department (Leverone 569) by 5:00 p.m. on Monday December 6, 1999 (fax: 847-467-1220).

More information may be shared at news://news.kellogg.northwestern.edu/class.dsd36.myerson.

Planned outline of topics:

1. Probabilities of events. Independence, conditional probabilities, conditional independence, basic techniques of simulation in spreadsheets. Main example: inference about a new employee's talent.
Assignment due Monday Oct 4: printed analysis of problem 1 at the end of Chapter 1.
[Chapter 1] [Spreadsheet for Chapter 1] (See also AWZ sections 4.1-4.2, 6.7.)

2. Discrete random variables. Simulation with inverse cumulatives, expected value, standard deviation, law of large numbers, 95% confidence intervals for the expected value from sample data, the expected value criterion in decision-making, value-at-risk and cumulative risk profiles. Main example: uncertainty about the number of competitors who will enter a new market. Assignment due Mon Oct 11: printed analysis of problems 1 and 2 at end of Chapter 2.
[Chapter 2] [Spreadsheet for Chapter 2] [more problems] (See also AWZ sections 4.3-4.6, 8.1-8.3.)

3. Random variables with continuous distributions. Normal distribution, central limit theorem, Lognormal distribution for growth rates, the EXP and LN functions, fitting Generalized lognormals for subjectively assessed quartiles. Main example: continuation of the example from week 2, but with cost and market size treated as continuous random variables. Assignment in chapter 3: do problem 3 (subjective assessments) by Mon Oct 18, do problems 1 and 2 (Lepton case) by Wed Oct 20.
[Chapter 3] [Spreadsheet for Chapter 3] (See also AWZ chapter 5.)

4. Joint distributions of multiple random variables. Covariance and correlation, using CORAND to simulate Multivariate Normals, linear combinations of random variables and diversified portfolios, introduction to Solver, subjective assessment of correlations, correlations of nonNormal random variables. Main example: portfolio diversification problem. Assignment due Nov 1: hand in printed answers to problems 1 and 2 for chapter 4, and prepare part C of "Superior Semiconductor" for class discussion.
[Chapter 4] [Spreadsheet for Chapter 4] (See also AWZ sections 3.7, 4.7-4.10, 15.8.3.)

5. Conditional expectation. Statistical dependence and formulaic dependence, the law of expected posteriors, introduction to simple regression models.
[Chapter 5] [Spreadsheet for Chapter 5].

6. Decision variables. Analysis of decision variables in simulation models, strategic use of information, the winner's curse, more on use and limitations of Solver. Also, an introduction to risk aversion: utility functions and certainty equivalents for a decision-maker with constant risk tolerance. First example: newsboy problem. Second example: bidding problems.
Assignment due Nov 17: hand in printed answers to problem 1 in chapter 5 and problem 1 in chapter 6. Also prepare problems 2 and 3 in chapter 6 for class discussion on Nov 17. The Valdez Wilderness and Bates (B) bidding problems will be discussed in class discussion on Mon Nov 22.
[Chapter 6A] [Spreadsheet for Chapter 6A] [Chapter 6B:bidding notes] [Bidding spreadsheets]

7. Time series forecasting models. Main examples: planning when to terminate a project, finding a maximal debt burden that can be supported with high probability by uncertain income stream, logBrownian motion models in finance. Cases to prepare for class discussion: Global Sports for Mon Nov 29.
[Chapter 7A] [Spreadsheets for Chapter 7A] [More cases (texts)] [More cases (spreadsheets)]

(8. Risk sharing. Optimal risk sharing among risk-averse investors, arbitrage pricing theory and implicit market probabilities. Main examples: optimal portfolio planning, risk sharing.)

Some texts on decision analysis and simulation in spreadsheets:

  • Data Analysis and Decision Making, by S. Christian Albright, Wayne L. Winston, and Christopher Zappe (Duxbury, 1999)
  • Quantitative Risk Analysis, by David Vose (Wiley, 1996).
  • Making Hard Decisions: An Introduction to Decision Analysis, by Robert T. Clemen (Duxbury, 1996).
  • Simulation Modeling Using @Risk, by Wayne L. Winston (Duxbury Press, 1996).
  • Practical Management Science, by Wayne L. Winston and S. Christian Albright (Wadsworth 1997), chapters 8-14.
  • Spreadsheet Modeling and Decision Analysis, by Cliff T. Ragsdale (Course Technology, 1995), chapters 12-16.
  • Insight.xla, by Sam L. Savage (Brooks/Cole, 1998).
  • Beyond Value at Risk: The New Science of Risk Management, by Kevin Dowd (Wiley, 1998).

For more references see the INFORMS Bookstore's decision analysis book list and Gloriamundi.org's risk management book list.

Course description:

DS-D36 is an accelerated and extended version of DS-D33, Kellogg's introductory Decision Sciences course, which covers basic topics in probability and statistics. DS-D36 is offered as an alternative to DS-D33 for students who have had some previous study in probability and statistics. In particular, students in DS-D36 should have seen expected values, standard deviations, conditional probabilities, and Normal distributions in some previous course.

In DS-D36, students review the basic ideas of probability and statistics while learning to apply them to complex managerial problems with advanced spreadsheet models. The emphasis is on practical applications of probability for analysis of decisions under uncertainty, using Monte Carlo simulation in Microsoft Excel.

The required text materials for DS-D36 will be distributed in class and at this site. A recommended supplementary text is Data Analysis and Decision Making with Microsoft Excel, by Christian Albright, Wayne Winston, and Christopher Zappe (Duxbury Press, 1999), abbreviated below as AWZ.

All the cases that will be studied in DS-D36 this year are designed to develop students' skills for advanced spreadsheet analysis with Excel. This emphasis on using Excel is new and experimental in DS-D36 this year, and most case materials for this course have been used previously only in the advanced elective course DS-D50. Students who do not take DS-D36 this fall may consider taking DS-D50 in Winter 2000.

Two add-ins for Excel, simtools.xla and formlist.xla, have been developed for use in this course. These add-in files are available with documentation and installation instructions at http://www.kellogg.northwestern.edu/faculty/myerson/ftp/addins.htm.

First assignment: Download and install the Excel add-ins simtools.xla and formlist.xla. Then download the first reading "Introduction to Probability and Simulation in Spreadsheets" and the corresponding workbook SALES.XLS. Read at least the first 7 pages (to the end of section 1) before the first class.