Pages

28 September 2022

Rapidly Detecting and Correcting Degradation of Military Supply Distribution Performance

James R. Broyles, Kenneth J. Girardini, Jason Mastbaum, Marc Robbins, Patricia 

Army units worldwide depend on a complex network of distribution centers, managed primarily by the Defense Logistics Agency, to support equipment readiness and sustainability. Rapid and reliable logistics distribution support is especially important for U.S. Army forces deployed into theaters of operations. There are many factors that can cause performance changes affecting the distribution timeliness to the Army. Currently, distribution problems are detected manually and reactively by Army units once these problems start to affect equipment readiness. This report describes (1) algorithms developed by the authors that monitor the logistics distribution system and automatically detect distribution problems (or potential distribution problems) that might affect equipment readiness, and (2) data visualizations developed by the authors that assist Army managers and analysts to determine the root causes and potential corrective actions related to the detections. The report also provides several case studies illustrating the algorithms' effectiveness.

Key Findings

Current Army distribution metrics are lagging indicators of problems because they focus on requisition wait time, which requires the receipt of the shipment, and require manual monitoring to detect problems.

In several historical case studies of distribution performance degradation, the detection algorithms and metrics could have automatically detected actual or potential distribution problems several months prior to when they were realized by the Army units.

Recommendations

Implement the detection algorithms and visualizations in an Army analytics platform for continued use and to inform corrective actions.

Expand the metrics beyond requisition wait time and provide open shipment data in the analytics platform so that the algorithms can identify potential distribution problems earlier.

No comments:

Post a Comment