Sciweavers

Free Online Productivity Tools i2Speak i2Symbol i2OCR iTex2Img iWeb2Print iWeb2Shot i2Type iPdf2Split iPdf2Merge i2Bopomofo i2Arabic i2Style i2Image i2PDF iLatex2Rtf Sci2ools

9

ICDCSW
2006
IEEE

favoriteEmaildiscussreport

133views Computer Networks» more ICDCSW 2006»

Improve Searching by Reinforcement Learning in Unstructured P2Ps

13 years 10 months ago

Improve Searching by Reinforcement Learning in Unstructured P2Ps

Download www.cse.fau.edu

— Existing searching schemes in unstructured P2Ps can be categorized as either blind or informed. The quality of query results in blind schemes is low. Informed schemes use simple heuristics that lack the theoretical background to support the simulation results. In this paper, we propose to improve searching by reinforcement learning (RL), which has been proven in artiﬁcial intelligence to be able to learn the best sequence of actions in order to achieve a certain goal. Our approach, ISRL, aims at locating the best path to desired ﬁles at low cost. It explores new paths by forwarding queries to randomly chosen neighbors. It also exploits the paths that have been discovered to reduce the cumulative query cost. Two models of ISRL are proposed: the basic ISRL for ﬁnding one desired ﬁle, and MPISRL for ﬁnding multiple desired ﬁles. ISRL outperforms existing searching approaches in unstructured P2Ps by achieving higher query quality with less query trafﬁc. The experimental r...

Xiuqi Li, Jie Wu

Real-time Traffic

Blind Schemes | Cumulative Query Cost | Distributed And Parallel Computing | ICDCSW 2006 | Unstructured P2ps |

claim paper

Related Content

» A dynamic routing protocol for keyword search in unstructured peertopeer networks

» Isrl Intelligent Search by Reinforcement Learning in Unstructured Peertopeer Networks

» Reinforcement Learning for QueryOriented Routing Indices in Unstructured PeertoPeer Networ...

» Reinforcement Learning for P2P Searching

» Exploiting Multiple Secondary Reinforcers in Policy Gradient Reinforcement Learning

» Covariant Policy Search

» A Distributed Reinforcement Learning Approach to Pattern Inference in Go

» Improving reinforcement learning function approximators via neuroevolution

» High speed obstacle avoidance using monocular vision and reinforcement learning

Post Info
More Details (n/a)

Added	11 Jun 2010
Updated	11 Jun 2010
Type	Conference
Year	2006
Where	ICDCSW
Authors	Xiuqi Li, Jie Wu

Comments (0)