Abstract:

The problem of combat command decision group partition is analyzed based on the correlation degree of combat tasks. The previous measures of group partition need to preestablish the number of decision groups, and not make the best use of the correlation degree transitivity between the members of a group. The cluster algorithm based on the correlation degree of command decision units’ intensions is designed. The distance between intentions is proposed, and the method of setting an adaptive threshold is given. The problem of group partition is transferred to the problem of getting connected partial graphs from a weighted complete graph in graph theory, and the solving algorithm is given. The feasibility and validity of this method are proved by an example.