Capabilities-based force aggregation using random sets

Force aggregation is one of the most important functionalities of a situation analysis system. In order to reduce the amount of information displayed for an analyst, it is vitally important to cluster information that belongs together and display the aggregated information for the cluster instead of all the original objects. In order to do this, two different kinds of algorithms are necessary. First, we must have a method for grouping objects that belong together. This problem is often referred to as clustering or association; it is a variant of the NP-complete graph coloring problem. Second, a group of objects that belong together must be classified. There have been some methods proposed for doing this. All of the present alternatives for aggregation rely on doctrinal information. However, in the new kind of situations that face us, it is increasingly likely that we will meet organizations that do not follow a strict doctrine in their organization. Instead, they will use task-forces or ad-hoc forces that are organized to solve a specific objective. Here, we present formalism for doing classification of task-forces based on less amount of doctrinal knowledge. The kind of doctrinal knowledge required by the approach suggested here is similar to the one needed to put together task-forces for solving a specific mission, i.e., it is capabilities oriented. Using random set theory, we describe several different ways of force aggregation and present results from experiments performed with them. User interaction could be used to further enhance the method presented here.