A Multi-:Layer Peer to Peer Information Retrieval Simulation Environment
The core problem of Peer-to-peer based information retrieval systems is to efficiently route queries to desired information sources. Given the assumption that agents are only in possession of limited (or no) non-local information, it remains a challenging problem to perform informed searches on such systems.This work is one of the first attempts to address this issue from an organizational perspective. The agent organization of a peer-to-peer information retrieval system includes both the underlying topological network and the coordination mechanism among agents in the process of query routing. Efforts from both aspects have been made to achieve better information retrieval performance and improve system throughput. Specifically, it includes the following components: (1)A reorganization process to form implicit agent clusters. (2)A hierarchical agent organization formation algorithm; (3)A learning based adaptive self-organizing approach.
