Expert systems for computerised wargames

Computerised wargames have emerged as an important tool to train the field commanders in a cost-effective manner, because of its ability to incorporate all the mechanics and vagaries of warfare, and at the same time reduce the cost and complexity of stage-managing such training. However, while trying to convert tactical combat rules into a computerised system within the rigid limitations of software semantics, the very essence and dynamics of the phenomenon of warfare as manifested in field are likely to be lost. Such loss of critical aspects could make the output of the system unrealistic, which in turn may compromise its training value. Besides this, to develop a system based on software directly translated from conventional rules, one has to clearly define the phenomenon of warfare at extremely high resolution and accuracy. The process of defining a highly uncertain phenomenon like warfare at such high resolutions and thereafter, framing extensive rules for all the possibilities can make the system extremely complex and therefore unmanageable in many ways. This paper attempts to simplify this problem by proposing a simpler and better technique using an algorithm based on fuzzy logic. Its basic advantage over conventional systems is that it has the inherent potential to handle even highly complex phenomenon like warfare in a fundamentally simple manner. Such potential makes it capable of handling higher level of details and still contain the complexity of the software within manageable limits. Additional details would also make the system more accurate and realistic. Introduction of flexible models, like the proposed one, would definitely help improve the realism of the outcome generated in computerised wargames, thereby enhancing their training value.