Utilizing Patterns in US Military Interventions to Improve Logistics Decision Making
Abstract
The field of political science has had difficulty predicting where the next conflict will occur using strictly quantitative methods. However once a conflict does occur, there seems to be some logical variables such as oil or level of democracy that contribute to the US's willingness to commit military forces abroad. How these variables relate and interact in determining the US decision of whether or not to enter a conflict is a difficult matter. No known traditional linear model to predict US conflict decisions has been formulated. This research proposes a list of variables that might impact intervention decisions and puts forth a neural network approach to analyzing the underlying interactions present in existing conflict data. This method explores the interactive and possibly non-linear nature of conflict decision making. The results indicate that a reasonably small number of variables can be used to predict when the US might enter a conflict. Finally, the results of this analysis are then applied to a logistics location problem in order to show how eliminating some degree of uncertainty might significantly improve inventory positioning decisions.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jan 17, 2002
- Accession Number
- ADA398515
Entities
People
- John E. Bell
Organizations
- Auburn University