The organization and collaborative protocols of agent societies are becoming increasingly important with the growing size of agent networks. Particularly, in a multi-agent based content sharing system, a flat, peer-to-peer (P2P) agent organization is not the most efficient organization for locating relevant agents for queries. This work develops and analyzes a hierarchical agent group formation protocol to build a hybrid organization for large-scale content sharing system as well as a content-aware distributed search algorithm to take advantage of such an organization. During the organization formation process, the agents manage their agent-view structures to form a hierarchical topology in an incremental fashion. The algorithm aims to place those agents with similar content in the same group. We evaluate the system performance based on TREC VLC 921 datasets. The results of the experiment demonstrate a significant increase in the cumulative recall ratio (CRR) measure compared to the flat agent organization and structure.
[1]
Victor R. Lesser,et al.
A Multi-Agent Approach for Peer-to-Peer Based Information Retrieval System
,
2004,
AAMAS.
[2]
Christopher R. Palmer,et al.
Generating network topologies that obey power laws
,
2000,
Globecom '00 - IEEE. Global Telecommunications Conference. Conference Record (Cat. No.00CH37137).
[3]
James C. French,et al.
Comparing the performance of database selection algorithms
,
1999,
SIGIR '99.
[4]
David Hawking,et al.
Overview of TREC-7 Very Large Collection Track
,
1997,
TREC.
[5]
Edith Cohen,et al.
Associative search in peer to peer networks: harnessing latent semantics
,
2003,
IEEE INFOCOM 2003. Twenty-second Annual Joint Conference of the IEEE Computer and Communications Societies (IEEE Cat. No.03CH37428).
[6]
James P. Callan,et al.
Experiments Using the Lemur Toolkit
,
2001,
TREC.