Framework for optimizing intelligence collection requirements

In the military, typical mission execution goes through cycles of intelligence collection and action planning phases. For complex operations where many parameters affect the outcomes of the mission, several steps may be taken for intelligence collection before the optimal Course of Action is actually carried out. Human analytics suggests the steps of: (1) anticipating plausible futures, (2) determining information requirements, and (3) optimize the choice of feasible and cost-effective intelligence requirements. This work formalizes this process by developing a decision support tool to determine information requirements needed to differentiate critical plausible futures, and formulating a mixed integer programming problem to trade-off the feasibility and benefits of intelligence collection requirements. Course of Action planning has been widely studied in the military domain, but mostly in an abstract fashion. Intelligence collection, while intuitively aiming at reducing uncertainties, should ultimately produce optimal outcomes for mission success. Building on previous efforts, this work studies the effect of plausible futures estimated based on current adversary activities. A set of differentiating event attributes are derived for each set of high impact futures, forming a candidate collection requirement action. The candidate collection requirement actions are then used as inputs to a Mixed Integer Programming formulation, which optimizes the plausible future mission state subject to timing and cost constraints. The plausible future mission state is estimated by assuming that the Collection Requirement Actions can potentially avert the damages adversary future activities might cause. A case study was performed to demonstrate several use cases for the overall framework