Understanding EBO: model abstraction and achieving a favorable endstate

The planning, execution and assessment of Effects-Based Operations present many fundamental modeling problems. Futhermore, an understanding of adversary reactions and the causal linkages to our actions are critical parts of the problem of understanding whether objectives have been met. This paper will report on some recent work at ALPHATECH aimed at developing technology that will help enable understanding of Effects-Based Operations, and postulate some directions for future research. Under the DARPA Endstate program, ALPHATECH has been examining the problem of understanding the effects of operations on multiple, complex and highly interconnected networks within a nation-state's infrastructure. The unifying technical concept of much of the Endstate work is model abstraction. Endstate has considered two basic forms of model abstraction in connecting models of different systems. The first is reduced-order modeling, an inductive approach to modeling sub-problems in variable spaces of reduced dimensionality. The second form of model abstraction is deductive, with geospatial or timescale decomposition leading to hierarchies of related models. Model-based abstraction of network facilities and links, together with constraining physics, as well as the priorities and objectives of embedded controllers or coordinators are studied for networks of interest.