• Summary
• Related Documents
  • Active Bibliography
  • Co-citation
• Version History

Abstract:

Abstract universally being identified from many different occupations (e.g. civil aviation, military, etc.). Our research in Distributed Artificial Intelligence (DAI) has lead us to propose an innovative solution to the team concept training issue. We present the implementation details of a multiagent prototype modelling crew interpersonal communication network during the completion of a military mission aboard the anti-submarine aircraft AURORA. Based on the ACTOR computational model, this prototype is implemented on a distributed memory parallel machine, enabling several autonomous knowledge based systems and dynamic objects, executing simultaneously, to freely interact to solve complex problems. An appreciable gain in performance, attributed to parallelism, is added to the power of knowledge based systems applied to specific aspects of military operations. We have been studying systems which enable relatively independent heterogeneous agents to cooperate in solving specific problems that require their combined abilities (known as multi-agent systems). The parallel between our research and team training quickly became evident. By simple analogy, teams can be considered as being multi-agent systems (and vice-versa). Caution is advised in making this analogy as the general multi-agent paradigm is a concept still in its infancy with many intricate problems yet to be solved. It is nevertheless possible to define certain classes of problems for which solutions are within the realm of present technology. One such class of problems is created when applying the multi-agent paradigm to certain aspects of military teams or crews. We show that complex problems such as the completion of military mission can be modelled within a multi-agent environment. We briefly discuss the numerous military applications of such modellization, in particular, a new concept in simulation training that we call Virtual Role Playing.

Citations