Data Driven Multivariate ARIMA Model for Test Queue Predictions and Capacity Planning

Capacity Planning (CP) is an integral part of the Sales and Operations (S&OP) plan for forecasting resource utilization and availability needed to meet demands. It is used by industries who develop and manufacture products being delivered to customers. Manufacturers are regularly assessing CP methods to increase on time delivery, maintain reputation, and build revenue. Throughout several industries utilizing reliable data and accurate predictive methods are consistent key contributing factors for improving CP. Historically, CP derived from general algorithms (GA) applied to expert judgment, what if analysis, and standards. With an increase in data availability and modeling tools, CP has evolved to using data driven models with complex predictive algorithms. This research aims to introduce a data driven multivariate ARIMA model for CP of antenna Nearfield Range (NFR) test cells. There is no research to date that utilizes ARIMA to forecast antenna NFR cycle time for CP. Antenna nearfield range testing is a highly complex operation in antenna manufacturing and its test cycle time consists of many variables. It is advantageous to identify the contributing variables of antenna nearfield range cycle time and categorize them accurately to predict and plan capacity. The research starts with a detailed review of other research and work pertaining to data driven models, forecasting methodologies, and ARIMA to develop a baseline of prior research and limitations. The literature review also includes details on how forecast models, data driven models, and accurate CP impacts revenue, delivery and reputation. After developing a baseline of prior research, the praxis details the methodology of creating a multivariate ARIMA model to forecast NFR cycle time data. The ARIMA forecasted cycle time data is compared to two methodologies of expert judgment forecasted cycle time data to test three quantitative hypotheses. The two forecast methods are compared to determine which is more accurate in forecasting, which yields the highest utilization, and which results in the best on time delivery. The dependencies, limitations, and weaknesses of each method are compared in the results and conclusion. Lastly this research incorporates a pseudo manufacturing cell using a queuing model to mimic real life application of incorporating ARIMA into an everyday system for CP.

Author

• Walls, Deidra D

Language

• en

Keyword

• Antenna
• ARIMA
• Range
• Cycletime
• Nearfield
• Capacity

Date created

• 2026

Type of Work

• Dissertation

License

• All rights reserved

Rights statement

• In Copyright

GW Unit

• Engineering Management

Degree

• D.Engr.

Advisor

• Emmons, Debra L

Committee Member(s)

• Blackford, Joseph P
• Bierce, Jonathan

Persistent URL

•  https://scholarspace.library.gwu.edu/etd/tq57ns144