An algorithm for min-cut density-balanced partitioning in P2P web ranking

Abstract

© Springer International Publishing Switzerland 2015. In P2P-based PageRank computing, each computational peer contains a partitioned local web-link graph and its PageRank is computed locally. Then, collaborative web ranking between any two peers will be proceeded iteratively to adjust the web ranking until converge. In this paper, the problem of partitioning web-link graph for web ranking in P2P is formulated as a minimal cut-set with density-balanced partitioning. Then, an efficient algorithm called DBP-dRanking is proposed to address such problem. The algorithm can solve the problem with computational complexity of a polynomial function to the web-link graph size. The results also confirm that the proposed algorithm can reduce the ranking error by partitioning web-link graph and perform faster than two other algorithms.