A generic search strategy for large-scale real-world networks

Authors:
Yuichi Kurumida;Tsukasa Ogata;Hirotaka Ono;Kunihiko Sadakane;Masafumi Yamashita
Affiliations:
Kyushu University, Fukuoka, Japan;Kyushu University, Fukuoka, Japan and Fujitsu Network Technologies Limited;Kyushu University, Fukuoka, Japan;Kyushu University, Fukuoka, Japan;Kyushu University, Fukuoka, Japan
Venue:
InfoScale '06 Proceedings of the 1st international conference on Scalable information systems
Year:
2006

Citing 4
Cited 3

Randomized algorithms

Randomized algorithms
A tight upper bound on the cover time for random walks on graphs

Random Structures & Algorithms
Search in a Small World

IJCAI '99 Proceedings of the Sixteenth International Joint Conference on Artificial Intelligence
Statistical mechanics of complex networks

Statistical mechanics of complex networks

A Device Search Strategy Based on Connections History for Patient Monitoring

IWANN '09 Proceedings of the 10th International Work-Conference on Artificial Neural Networks: Part II: Distributed Computing, Artificial Intelligence, Bioinformatics, Soft Computing, and Ambient Assisted Living
Indexing Network Structure with Shortest-Path Trees

ACM Transactions on Knowledge Discovery from Data (TKDD)
Forest search: a paradigm for faster exploration of scale-free networks

ISPA'06 Proceedings of the 4th international conference on Parallel and Distributed Processing and Applications

Quantified Score

Hi-index	0.00

Visualization

Abstract

We consider the following situation for a given large-scale network: Starting from an initial node we move to its neighbor node and repeat that until reaching a target node. How fast can we do this without any global topological information? This problem is considered "searching networks", and several approaches have been proposed. In this paper, we present a general framework of search strategies, under which all of these existing approaches can be formalized. Our framework characterizes random search strategies by the following three parameters: memory for previously-visited nodes, look-ahead property and transition probability. Through computational simulations for large-scale networks with small-worldness and scale-freeness, we investigate the relationship between the effect of parameters of the strategies and the coefficients of networks such as the clustering coefficient. The comparison result provides a guideline to obtain good parameters of the strategies according to the diameter and the clustering coefficients of networks.