BI Ran, LI Jian-zhong, CHENG Si-yao. (ε,δ)-approximate Top-k query processing algorithm in wireless sensor networks[J]. 2011, 32(8): 45-54.DOI:
无线传感器网络(ε,δ)-近似Top-k查询处理算法
摘要
提出一种适合于任意数据分布的基于抽样的(ε
δ)–近似Top-k查询处理算法。其中
ε≥0和0≤δ<1分别是相对误差界和失败概率界。理论分析表明
对于任意ε≥0和0≤δ<1
该算法返回的查询结果的相对误差界大于ε/(1+ε)的概率小于δ。于是
该算法可以达到任意精度。同时
还给出了支持近似Top-k查询的优化的抽样算法
并通过节点上的数据过滤技术来减少通信能量的消耗。理论分析和仿真结果表明
提出的算法能量消耗低并且计算复杂度低。
Abstract
A sampling based approximate Top-k algorithm was proposed that is adaptive for any data distribution.δ≥0 and 0≤δ<1 are respectively relative error bound and failure probability bound.The theoretical analysis demonstrates that for any δ≥0 and 0≤δ<1 the probability that the relative error bound of the results returned by this algorithm is larger than ε/(1+ε) is less than δ.So the proposed algorithm can reach arbitrary precision.Furthermore
an optimal sampling algorithm was proposed that supported the approximate Top-k query
and through the technique of data filtering the en-ergy consumption of communication was reduced.Theoretical analysis and simulation show that the proposed algorithm is efficient and consumes little energy.