Complexity-based targeting: new sciences provide effects

Air & Space Power Journal, Spring, 2003 by Col Robert W. Freniere, Cmdr John Q. Dickmann, Cmdr Jeffrey R. Cares

THROUGHOUT THE MODERN history of bombardment, targeting philosophies have remained deeply rooted in industrial-age mind-sets and mechanistic, linear analyses of systems as engineered entities. As a result, in most significant bombing campaigns, targets have been classified by their physical attributes alone. For example, in the "serial bombing" philosophy of World War II, aircraft attacked large sets of physical targets sequentially. (1) Contemporary targeting philosophy--the "parallel warfare" employed during the Gulf War--advocates attacking targets with more simultaneity yet still focuses almost exclusively on their physical attributes and their engineered physical interactions. (2) In general, these targeting constructs are exceedingly inefficient, requiring inordinate amounts of "inputs" (tonnage of bullets and bombs, amounts of information warfare [IW], etc.) often not justified by or traceable to observed "outputs" (effects). Since the end of Operation Desert Storm, bombing campaigns have evolved in co ncept toward an objective of having specific effects on the enemy and his systems; in practice, planners still choose targets based upon engineering analyses of physical systems and physical interactions inside those systems. Little has changed.

Recent research asserts that the American military has historically misunderstood the systemic nature of targets. (3) Targeting has remained inefficient and unpredictable because most targets of military value are elements in complex adaptive systems, which behave according to a radically different operating dynamic than do mechanistic systems. An evolving body of scientific work, based on understanding the emergent behaviors of large collections of interacting entities, describes the behavior of these systems. Although this body of work is collectively referred to as the "new sciences," this article uses the terms complexity theory or complex adaptive systems theory. Whereas industrial-age Newtonian analysis focuses on classifying targets according to their physical nature, complexity theory allows targeteers to focus on how targets interrelate, particularly in nonphysical ways. Complexity-based targeting emphasizes and exploits the characteristics of complex adaptive systems.

Theory of Complexity-Based Targeting

Two concepts from complexity theory underpin complexity-based targeting: complexity and entropy. Complexity is a measure of the degree to which a system contains large numbers of interacting entities with coherent behavior. Notionally, one can measure complexity from a value of zero to some maximum number. Zero complexity indicates a completely simple system; few entities have either minimal or no interactions. Generally, one can account for the behavior of such a system with a simple set of equations or a short description--for example, contemporary military combat models, replete with attrition equations. (4) Entropy, on the other hand, is a measure of the amount of work lost in a system due to destructive forces such as friction or interference. One can measure entropy on a scale from zero to one--zero indicating a completely linear system that loses no work and behaves predictably. Maximum entropy designates a completely chaotic system that loses all work and behaves randomly.

As the number of possible interactions in a system increases, entropy increases--as does the number of coherent behaviors. When the system becomes more complex, predicting specific events becomes more difficult, describing what is occurring in the system takes longer, and making mathematical calculations becomes more involved. Complexity increases to a point that the interacting entities and groups of entities become too numerous and interfere with each other, and the aggregate behavior of the system becomes more random. As interference increases, so does entropy, causing complexity to fall to zero because the system's aggregate behavior becomes simple (i.e., all behaviors can cancel each other out, and one can usefully describe the system at some higher scale in much the same way one can describe the temperature of a gas without listing the temperature of each molecule).

Between the extremes of complete linear simplicity and complete chaotic simplicity lies a wide range of complex systems, including those containing most targets of military significance. Examples include electrical distribution grids, transportation networks, communications architectures, command and control organizations, naval missile exchanges, and ground combat. We call such examples complex adaptive systems because they meet our criterion of having a large number of interacting entities that can adapt to their environment as it changes (fig. 1). (5)

Complex adaptive systems are difficult to defeat because they have many groups of entities with coherent behavior. In a military context, as some entities are attacked, others change their behavior or alter their interactions, allowing the larger system to adapt. For example, if bridges in a road network are destroyed, maneuver forces will find other means--such as alternate routes, temporary bridges, or river fords--to accomplish their mission. Complexity-based targeting seeks to prevent a complex adaptive system from using its attributes and mechanisms in response to an attack.