Time Series Modeling of Army Mission Command Communication Networks: An Event-Driven Analysis.

摘要

We examine the communication time series of a fully-networked Army coalition command and control organization. The communications and scenario- event data described in this paper were collected at a two-week U.S. Army simulation-based training event. The Mission Command Battle Laboratory at Fort Leavenworth, Kansas, conducted a joint experiment/training exercise focused on the operations of the mission command staff composed of a U.S. Division headquarters (n=51) and subordinate U.K. Brigade headquarters (n=28). The network architecture and digitized nature of the event allowed examination of staff communications in a distributed, network-enabled coalition environment. The participants were active duty Soldiers and officers, organic to their military unit. We used time series analysis to predict the communications record based on an external work variable of the number of important scenario events occurring across time. After taking into account structural features of the time series, we examined the remaining variability in email and phone communications. We found that the exercise scenario events were not a significant predictor of the Divisional communications, which were best fit by an auto-regressive model of order 1. The occurrence of scenario events, however, did predict the Brigade communications time series, which were well fit by a lag dependent variable model. These results demonstrate that Brigade communications responded to and could be predicted by battlefield events, whereas the Division communications were only predicted by their own past values. These results highlight the importance of modeling environmental work events to predict organizational communication time series and suggest that network communications are perhaps increasingly dependent upon battlefield events for lower echelons of command closer to the tactical edge.