A Predictive Model Approach to Military Construction Prioritization and Facility Sustainment

The Department of the Navy (DoN) of the United States faces significant inefficiencies and productivity losses in maintenance and construction backlogs. The DoN's current decision models are insufficient for accurately addressing sustainment costs such as long-term maintenance, structural integrity, energy, and modernization costs. This praxis will develop a decision aid for naval leadership to optimize decision-making by determining the most critical facilities, maximizing current military infrastructure assets, and minimizing costs using existing available data. This data-driven model will incorporate previously ignored quantifiable data to improve project funding by re-evaluating how criteria are calculated at the installation level and weighted at the service component level. The data-driven model is developed by using ensemble machine learning methods to gain the advantages of more than one algorithm. A Pareto frontier multi-objective optimization algorithm is used to optimize cost and other parameters associated with the predictive model. The results show the ensemble learning used along with boosting algorithms gives better results as compared to other individual models achieving an accuracy of 93.61%. The predictive scores were obtained to analyze the importance of the features. The model developed is the optimized predictive data-driven model and can be utilized in real scenarios during the prioritization process.

Author

• VILLALTA SALAMANCA, KRIZTOFER

Language

• en

Date created

• 2023

Type of Work

• Dissertation

License

• All rights reserved

Rights statement

• In Copyright

GW Unit

• Engineering Management

Degree

• D.Engr.

Advisor

• KAMP, JOHN C

Committee Member(s)

• JONES, MICHAEL C
• ETEMADI, AMIRHOSSEIN

Persistent URL

•  https://scholarspace.library.gwu.edu/etd/vx021f95d