Date on Master's Thesis/Doctoral Dissertation

5-2009

Document Type

Master's Thesis

Degree Name

M. Eng.

Department

Industrial Engineering

Committee Chair

Depuy, Gail W.

Subject

United Parcel Service; Aeronautics, Commercial--Freight--Data processing; Forecasting--Data processing

Abstract

This thesis develops a forecasting model to predict six different volume measures on a weekly and daily basis for UPS-Supply Chain Solutions (UPS-SCS). The volume measures are used by UPS-SCS to develop business plans, operation plans, and staffing plans. Four different forecasting methods are used to evaluate each volume metric. Moving average, single exponential smoothing, double exponential smoothing and Winter’s additive model are the four forecasting methods that are used to generate forecasts. From the statistical evaluation measures and graphs of the data, decisions of which forecasting method for each volume measure can be made. The demand pattern of the data set will influence which of the forecasting methods should be selected. UPS-SCS currently has a weekly forecasting method but it is not complex or dynamic. Two different case studies of actual data sets showed that the UPS-SCS Weekly Forecasting tool generated 10% more accurate forecasts compared to the current method that UPS-SCS uses. Based on the two different case studies the UPS-SCS Weekly Forecasting tool generated forecasts that were 90% effective or better and the current method that UPS-SCS uses generated forecasts that were 80% effective.

Share

COinS