作者: Lu Guan , Ying Wang , Wenjing Li , Congxian Yan
DOI: 10.1109/CNSM.2013.6727837
关键词:
摘要: Adaptive active diagnosis method is widely adopted for fault in networks. In diagnosis, appropriate probes are selected sequentially and made by inference from results of probes. It very important to select with low cost less impact on network performance. However, the selection most informative set limited an NP-hard problem. The computational complexities existing probe algorithms still too high large scale this paper, a lemma about mutual information provided proved based property conditional entropy. Then approximate derived introduced compute probe. With efficient algorithm proposed. At last, efficiency effectiveness proposed verified through simulation.