LRS: A Novel Learning Routing Scheme for Query Routing on Unstructured P2P Systems

摘要: Query routing is a fundamental problem in unstructured Peer-to-Peer systems. Recently, researches this area have focused on methods based query-oriented indices. These use the historical information of past queries and query hits to build local knowledge base per peer, which represents user’s interests or profile. Existing approaches represent profile only by some statistics about they not addressed two difficult challenging problems: (i) bootstraping (ii) unsuccessful relevant peers search. Indeed, when peer selects an insufficient number from its base, it floods through network, badly affects efficiency effectiveness. To tackle these problems, we introduce novel Learning Routing Scheme (LRS). We implemented proposed scheme compared retrieval effectiveness with broadcasting (without learning) learning taken literature. Experimental results show that our carries out better than other ones respect accuracy.