Efficiently answering probability threshold-based sp queries over uncertain graphs

Ye Yuan*, Lei Chen, Guoren Wang

*此作品的通讯作者

科研成果: 书/报告/会议事项章节会议稿件同行评审

54 引用 (Scopus)

摘要

Efficiently processing shortest path (SP) queries over stochastic networks attracted a lot of research attention as such queries are very popular in the emerging real world applications such as Intelligent Transportation Systems and communication networks whose edge weights can be modeled as a random variable. Some pervious works aim at finding the most likely SP (the path with largest probability to be SP), and others search the least-expected-weight path. In all these works, the definitions of the shortest path query are based on simple probabilistic models which can be converted into the multi-objective optimal issues on a weighted graph. However, these simple definitions miss important information about the internal structure of the probabilistic paths and the interplay among all the uncertain paths. Thus, in this paper, we propose a new SP definition based on the possible world semantics that has been widely adopted for probabilistic data management, and develop efficient methods to find threshold-based SP path queries over an uncertain graph. Extensive experiments based on real data sets verified the effectiveness of the proposed methods.

源语言英语
主期刊名Database Systems for Advanced Applications - 15th International Conference, DASFAA 2010, Proceedings
155-170
页数16
版本PART 1
DOI
出版状态已出版 - 2010
已对外发布
活动15th International Conference on Database Systems for Advanced Applications, DASFAA 2010 - Tsukuba, 日本
期限: 1 4月 20104 4月 2010

出版系列

姓名Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
编号PART 1
5981 LNCS
ISSN(印刷版)0302-9743
ISSN(电子版)1611-3349

会议

会议15th International Conference on Database Systems for Advanced Applications, DASFAA 2010
国家/地区日本
Tsukuba
时期1/04/104/04/10

指纹

探究 'Efficiently answering probability threshold-based sp queries over uncertain graphs' 的科研主题。它们共同构成独一无二的指纹。

引用此