33:799:301 Lecture Notes - Lecture 2: Forecast Error, Exponential Smoothing, Bullwhip

• School: Rutgers University
• Department: Supply Chain Management
• Course: 33:799:301
• Professor: D Klepacki

Document Summary

Two kinds of demand: independent demand: demand that ultimate customer controls; need to gure out independent demand, dependent demand: once you know independent demand, you can easily determine dependent demand instantly. 2 basic forecasting techniques: qualitative: people"s opinions & intuition. Personal insight: go to most knowledgable/experienced people & ask for their opinion fast response, but can be unreliable. Jury of executive opinion: use a panel of experienced people to discuss forecast. Delphi method: get opinions of individuals remotely (online) more objective than jury of executive opinion. Sales force estimation: call sales department together & discuss forecasting. Customer survey: ask customers directly about product opinion: quantitative: numbers, mathematical models. Cause & effect: assumes one or more factors predict future demand: simple linear regression: dependent on one independent variable, multiple linear regression: uses 2+ independent variables & dependent variable (demand) Neither technique works goal is to minimize forecast error. Forecast error: mean absolute deviation, mean absolute percent error, mean squared error.