Forecasting Workload for Defense Logistics Agency Distribution

Abstract

The Defense Logistics Agency (DLA) predicts issue and receipt workload for its distribution agency in order to maintain adequate staffing levels and set proper rates for customers. Inaccurate forecasts lead to inaccurate staffing, subsequently leading to inaccurate pricing. DLA s current regression forecasting model is no longer adequate for predicting future workload for DLA Distribution. We explore multiple forecasting techniques and provide a methodology for selecting a model that is a viable and accurate alternative for DLA. Our methodology encompasses best-fit determination, a comparison of predictability through back-casting, and a sensitivity exercise to see reaction and stability of our selected models predictions. Finally, we compare our best performing model with the current regression model to see what would have been reported if our model had been used instead of the current model for recent Program Budget Review (PBR) cycles. Our results suggest that an auto-regressive integrated moving average (ARIMA) model used with critical assessment and managerial judgment offers a viable alternative to the current model for predicting distribution workload.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Dec 01, 2014
Accession Number
ADA619687

Entities

People

  • Aaron W. Chonko
  • Padraic T. Heiliger
  • Travis W. Rudge

Organizations

  • Naval Postgraduate School

Tags

Communities of Interest

  • Biomedical
  • Human Systems

DTIC Thesaurus Topics

  • Business Administration
  • Computational Science
  • Contingency Operations (Military)
  • Data Analysis
  • Delphi Method
  • Department Of Defense
  • Economic Analysis
  • Information Science
  • Logistics
  • Operations Research
  • Personnel Management
  • Regression Analysis
  • Spreadsheet Software
  • Supply Chain
  • Supply Chain Management
  • United States
  • Workload

Readers

  • Logistics and Supply Chain Management.
  • Regression Analysis.