Course of Action Analysis for the U.S. Army's 2035 Modernization
Abstract
The United States Army Futures Command (AFC) is comparing multiple modernization levels of various programs in order to prepare for military operations in 2035 and beyond. This thesis supports that effort by analyzing a data set of three modernization level possibilities for 68 programs. There are 2.78 nonillion possible combinations in this data set, and the current methodology requires an analyst to select a single combination, and then calculate the overall benefit score. The time to complete this calculation is over 1 second, so by using the current methodology it would require 8.8*10(exp 22) centuries to compute. Therefore, this thesis started with two objectives: (1) develop a faster method for calculating the benefit score for a singular combination of programs, and (2) develop a methodology for comparing multiple scores to each other at once. While in pursuit of the first two objectives, the programming Julia proved to be exponentially faster for singular calculations, resulting in the ability to view nearly 9,000 scores within one second, all while using CPU encoding. These faster calculations led to the development of a third objective: compare the speed and accuracy of a machine learning (ML) algorithm. The third objective resulted in speeds 40 times faster than the CPU model, but with a relative error of 1.3%.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jun 01, 2021
- Accession Number
- AD1150844
Entities
People
- David R. Black
Organizations
- Naval Postgraduate School