Decision Analysis and Multicriteria Decision Making
This course complements and extends both 95-760 Decision Making under Uncertainty and the two mini sequence in Management Science (90-722 & 90-760 or 90-755 & 90-775). Management Science I: Optimization) by addressing three additional topics in managerial decision making:
MCDM is a collection of methods for trading-off different alternatives’ performance on multiple conflicting objectives; methods discussed include weighted sum scoring models, swing weights, TOPSIS, DEA, AHP, and rank-based methods.
Decision Analysis is the prescriptive model for rationally maximizing subjective expected utility in the face of uncertainty; it is particularly powerful for dealing with sequential decisions, quantifying the value of information, assessing and incorporating subjective probabilities, and doing Bayesian updates of probabilities as new information becomes available.
Decision process considerations go beyond the paradigm of a single well-defined decision maker and mathematical method. Potential topics we could cover include industrial analytics, auctions, composite indicators (“US News & World Report” style ratings), balanced scorecards, “dashboards” of key performance indicators, group processes, and matching a decision method to the circumstances at hand.
The first two topics are skills-oriented. The objective is to empower students to apply the methods in professional practice. Students should leave the class able to apply the methods to routine problems, and with a foundation for further self-study when the decision’s stakes warrant use of more advanced variations of these methods.
The last component is more conceptual, providing brief exposure to issues one confronts when bridging between textbook methods and real world implementation. The twin learning goals are knowledge about the issues and ideas and developing critical thinking skills concerning the “meta decision” of when to employ which decision method, and the criteria relevant to making those judgments.
This course will normally be taken after either 95-760 Decision Making Under Uncertainty or the Management Science sequence (90-722 & 90-760), but the only formal prerequisite is fluency with algebra & Excel and knowledge of probability theory and distributions at a level obtained from having taken, or concurrently taking one of Heinz’ empirical methods courses. However, the course is pitched at a second-year level; first-year students who are comfortable with quantitative methods are welcome, but first-year students without a quantitative bent are encouraged to wait until their second-year.