Distributed data sources which employ taxonomy hierarchy to describe the contents of their objects are considered, and a super-peer-based semantic overlay network （SSON） is proposed for sharing and searching their d...Distributed data sources which employ taxonomy hierarchy to describe the contents of their objects are considered, and a super-peer-based semantic overlay network （SSON） is proposed for sharing and searching their data objects. In SSON, peers are dynamically clustered into many semantic clusters based on the semantics of their data objects and organized in the semantic clusters into a semantic overlay network. Each semantic cluster consists of a super-peer and more peers, and is only responsible for answering queries in its semantic subspace. A query is first routed to the appropriate semantic clusters by an efficient searching algorithm, and then it is forwarded to the specific peers that hold the relevant data objects. Experimental results indicate that SSON has good scalability and achieves a competitive trade-off between search efficiency and costs.展开更多
针对二进制一致性算法扩展性差、经验依赖性强的缺点,提出了一种 N 状态分布式二进制一致性算法。首先,基于Gossip算法的平均一致性思想和轮盘赌思想,更新无线传感器网络状态均值和当前状态均值的偏差程度,计算所有可能更新状态的初始...针对二进制一致性算法扩展性差、经验依赖性强的缺点,提出了一种 N 状态分布式二进制一致性算法。首先,基于Gossip算法的平均一致性思想和轮盘赌思想,更新无线传感器网络状态均值和当前状态均值的偏差程度,计算所有可能更新状态的初始概率分布;然后,利用遗传算法优化初始概率分布,得到准确率较高的最优概率分布。仿真结果表明,在相同状态个数条件下,本文所设计的算法具有更好的准确率和收敛时间。展开更多
基金The National Natural Science Foundation of China(No60573089)the Natural Science Foundation of Liaoning Province(No20052031)the National High Technology Research and Develop-ment Program of China (863Program)(No2006AA09Z139)
文摘Distributed data sources which employ taxonomy hierarchy to describe the contents of their objects are considered, and a super-peer-based semantic overlay network （SSON） is proposed for sharing and searching their data objects. In SSON, peers are dynamically clustered into many semantic clusters based on the semantics of their data objects and organized in the semantic clusters into a semantic overlay network. Each semantic cluster consists of a super-peer and more peers, and is only responsible for answering queries in its semantic subspace. A query is first routed to the appropriate semantic clusters by an efficient searching algorithm, and then it is forwarded to the specific peers that hold the relevant data objects. Experimental results indicate that SSON has good scalability and achieves a competitive trade-off between search efficiency and costs.
文摘针对二进制一致性算法扩展性差、经验依赖性强的缺点,提出了一种 N 状态分布式二进制一致性算法。首先,基于Gossip算法的平均一致性思想和轮盘赌思想,更新无线传感器网络状态均值和当前状态均值的偏差程度,计算所有可能更新状态的初始概率分布;然后,利用遗传算法优化初始概率分布,得到准确率较高的最优概率分布。仿真结果表明,在相同状态个数条件下,本文所设计的算法具有更好的准确率和收敛时间。
