基于局部乱序事件排序的复杂事件匹配方法

Complex event matching method based on local out-of-order event sorting

为处理乱序事件流上的复杂事件匹配问题,减少排序开销,在非确定有限状态自动机的基础上提出解析查询指导排序策略和匹配结果反馈排序策略。根据查询内容对事件流进行分解过滤,得到查询所需的子事件流,设计基于堆排序的子事件流合并算法。通过匹配过程中自动机的状态,设计算法对自动机状态转移相关的事件实例进行排序。在股票数据集上进行实验,验证了提出方法相较于现有方法在匹配性能上有显著提升。

To deal with the complex event matching problem on the out-of-order event stream and reduce the sorting overhead,a parsing query guidance ordering strategy and a matching result feedback ordering strategy were proposed on the basis of non-deterministic finite automaton.The event stream was decomposed and filtered according to the query content to obtain the required subevent stream,and a subevent stream merging algorithm based on heap sorting was designed.By matching the state of the automaton,the algorithm was designed to sort the event instances related to the state transition of the automaton.Experimental results on the stock dataset show that the proposed method has a significant improvement in matching performance compared with the existing method.