Search

Ambushes, in the form of improvised explosive devices (IEDs), have posed grave risk to targeted vehicles operating on supply routes in recent theaters of war. In fact, history shows that this is an enduring problem that US military forces will likely face again in the future. This paper introduces a fundamental reinforcement learning (RL) model for determining convoy schedules and route clearance assignments in light of attack costs on a transportation network subject to IED ambushes. The model represents opponent interaction by assuming dependence between attack probabilities and targeted traffic patterns. There are currently few analytical approaches for this problem in the literature, but RL algorithms offer opportunities for meaningful improvements by optimizing individual movements across an extended planning horizon, accounting for downstream attacker-defender interaction. To our knowledge this approach has not been pursued elsewhere; therefore, this paper introduces the RL methodology with a fundamental formulation and initial computational results which show meaningful performance improvements over a one-step, myopic decision rules. [ABSTRACT FROM AUTHOR]

Motivated by experiences of coalition military forces in Iraq and Afghanistan, we analyze the allocation of route clearance teams (RCTs) to search for and neutralize improvised explosive devices (IEDs) on roadways traveled by military convoys. We model the interaction of a single RCT and a single convoy operating over a given roadway. Our goal is to reduce IED risk by improving coordination between the RCT and the convoy. We treat the distribution of IEDs along the road prior to the passage of the RCT as a non-homogeneous Poisson process. The RCT finds and clears IEDs according to a Bernoulli process. Enemy forces may emplace (reseed) additional IEDs in the temporal gap between the RCT clearance sweep and the arrival of the convoy. IED risk is defined as the expected number of IEDs encountered by the convoy. We identify certain characteristics of optimal RCT schedules including: the shape of the IED intensity function and the speed of reseeding substantially dictate the RCT schedule that minimizes IED risk; the more rapid the IED reseeding, the more critical is conformance to the optimal RCT schedule; an RCT having inferior detection probability, by employing superior scheduling can sometimes reduce convoy risk more than a less well scheduled RCT with superior detection probability; a discretized highway version of the problem can be efficiently optimized. We also briefly discuss applications of this model to the more complex problem of allocating several RCTs to protect a number of convoys scheduled to travel on a network of highways. [ABSTRACT FROM AUTHOR]