Will a crop failure in a country cause such a level of instability to be considered a national security threat? Are certain troop deployments better for deterrence, or will they cause an unintended war?

The multitude of “what would happen if” questions keeps military planners up at night, and proves to be difficult to simulate. Now, BAE Systems may have the answer.

The Defense Advanced Research Projects Agency has sponsored a $4.2 million Phase 1 contract to develop software that will “aid military planners in understanding and addressing the complex dynamics that drive conflicts,” according to a statement by the contractor.

The mostly manual methods currently used to predict and prepare for every contingency “have failed” planners in several ways, according to Jonathan Goldstein, senior principal scientist of the autonomy, controls and estimations unit at BAE Systems.

The existing modeling tool is “too opaque” (planners don’t understand why the model gave a particular outcome), “too rigid” (new information can’t be added or errors fixed) and “too burdensome” (requiring a lot of effort to configure, update and use), Goldstein asserted.

All this leads to “extremely time-consuming and labor-intensive” planning work for officials trying to figure out how global dominoes may fall, said Chris Eisenbies, product line director of the autonomy, controls and estimation unit.

Enter: CONTEXTS. Causal Modeling for Knowledge Transfer, Exploration, and Temporal Simulation gathers information about a conflict from intelligence reports, government databases, news and social media to automate and construct models at a low burden.

Eisenbies promises the program will use “reasoning algorithms and simulations” to provide quick but deep insight on a range of different “political, territorial and economic tensions” that spur conflict. In an era when the Pentagon must consider threats as varied as Chinese hackers to Russian underwater drones to North Korea’s nuclear proliferation, planners must carefully study how rising tensions can play out.

The reasoning algorithms even work to determine “the root cause of some observed phenomena, like why is a particular region experiencing more extremist activity than another,” Goldstein said.

Andrew is a student in the class of 2020 at the University of Notre Dame.

Share:
More In Training & Sim