The development of new intelligent agents requires an interdisciplinary approach to programming. The initial challenge is to describe the desired agent behaviors and abilities without necessarily committing the agent development project to one particular programming language. What are the appropriate linguistic and logical tools for creating a top level, unambiguous, program-independent, and consistent description of the functions and behaviors of the agent? And how can that description then be translated easily into one of a number of program languages? This article provides a case study of the application of a simple Belief, Desire, and Intention (EDI) first order logic to a complex set of agent functions of a theoretical community of intelligent nano-spacecraft. The basic research was conducted at NASA-GSFC (Greenbelt), Advanced Architecture Branch, during the summer of 2001. The simple examples of applied BDI logic presented here suggest broad application in agent software development.
Truszkowski, Walt; Iyengar, Jagannathan V.; and Mills, Fred
"Describing intelligent agent behaviors,"
Journal of International Information Management: Vol. 10
, Article 10.
Available at: http://scholarworks.lib.csusb.edu/jiim/vol10/iss2/10