MACRON: An Architecture for Multi-agent Cooperative Information Gathering

The complexity of the modern information carrying landscape requires a sophisti cated view in which information is acquired rather than simply retrieved where the process must be dynamic incremental and constrained by resource limitations In this paper we describe a multi agent architecture called MACRON based on the Cooper ative Information Gathering CIG paradigm designed speci cally for such complex information gathering environments Top level user queries drive the creation of par tially elaborated information gathering plans resulting in the employment of multiple cooperative agents for the purpose of achieving goals and subgoals within those plans Agents in MACRON incorporate capabilities to exploit subproblem interdependen cies manage uncertainty inherent in multi agent search intelligently trade o solution quality for resource limitations and exploit or avoid redundancy as needed