Capturing Expertise: Some Approaches to Modeling Command Decisionmaking in Combat Analysis

No satisfactory technique yet exists in combat models for treating one major area of command decision making: the formulation and modification of concepts of operations, that is, the basic strategy around which all the activities of one side in a combat are organized. Existing methods of treating command decisionmaking in combat models are reviewed here: 1) human participation, 2) game theory, 3) optimization, 4) mechanical-statistical techniques, 5) controlled experimentation, and 6) expert systems. Three new approaches offer useful capabilities for improved treatment of concepts of operations. In the XMOD methodology, a concept of operations provided by human experts is expanded into program inputs and mechanical-statistical rules and is currently supported on the VECTOR series of corps-through-theater campaign models. MACRO, a highly aggregate theater model fitted to the results of detailed simulated corps campaigns, takes a more abstract view, by treating each commander as an explicit optimizer on input goals and constraints. In a third approach, rule-based systems are derived to simulate the decision-making behavior of gamers, based on the analysis of wargame histories and complemented by knowledge engineering techniques.