Modeling qualitative issues in military simulations with the RAND-ABEL language

This paper describes a methodology for developing combat simulations that may be readily tailored to specific study issues, and structured so as to treat both qualitative and quantitative variables using the natural language of military planners. The methodology was developed for the RAND Strategy Assessment System (RSAS) and exploits a new programming language called RAND-ABELTM. RAND-ABELTM can be understood and used by analysts who are not expert programmers, and allows key relationships to be expressed in a form that is more easily reviewable by non-analyst military officers. The RAND-ABELTM language has been used to build several combat simulations, which we refer to generically as Referee models. In Referee models, both the assessment and decision processes are rule-based, thereby making the models both transparent and flexible. Through the use of a selective RAND-ABELTM interpreter, the models may be modified even while it is running. This new approach to modeling has been successfully applied to several projects.