"Game Commander"-applying an architecture of game theory and tree lookahead to the command and control process

This paper describes an architecture for emulation of those portions of the higher-echelon command and control task which deal with planning and evaluation of courses of action (COA). The architecture described utilizes the technique of tree lookahead with game theory. Tree lookahead is a technique for computing optimal decisions. In chess it is now the preferred method in real time. It is precisely the chess application which motivates its adoption into an overall command and control architecture. Chess is the prototypical war game and the ways that chess players develop their strategies are not unlike the techniques used by commanders to reason about, plan, and evaluate courses of action to pursue in a combat situation. The paper examines implementation issues associated with this technique. It reviews some past work on developing this technique for controlling rotary-wing aircraft. It establishes how this architecture could be embedded in a virtual simulation computer generated forces host (e.g. "Modular Semi-Automated Forces" or ModSAF) that would assume other command and control functions and thus provide a broader context for command and control.<>