60 likes | 205 Vues
University of Chicago Graduate School of Business. Introduction to Computer Based Models Bus-36102-81 Mr. Schrage Spring 2003. Designing and Implementing Decision Support Systems.
E N D
University of Chicago Graduate School of Business Introduction to Computer Based Models Bus-36102-81 Mr. Schrage Spring 2003 Designing and Implementing Decision Support Systems
Choosing a Project Some Guidelines: Gulbenkian’s Theorem(Mr. 5% of the Persian Gulf) Work on the project with greatest bang/buck. Embezzler’s/Investment Banker’s Corollary: Position yourself near a large cash flow and claim a small percentage.
Making the Model AdaptableSeparate data from the model equations, Use the flat file/list style, Making spreadsheets scaleable. What kind of Approximations should we make: Functional Aggregation Temporal Cross-sectional Statistical DecompositionPassing the Snicker TestA Model should always have a feasible solution.
What Features Can We Safely Disregard? Uncertainty, e.g. in demand Data quality Equilibrium considerations Gamesmanship/arbitrage Incentives for players Multi-period aspects Integrality Complexity, important 2nd order effects What-If vs. What’s Best
Implementation Issues Validation Are my approximations too gross? Verification Are there bugs in my model?
Other Issues Should we behave nonoptimally? Carry inventory Credit granting Bluffing When your model and the RW disagree, who is correct? Marketing the model, Getting commitment Report formats.