摘要
主动数据库中的规则是由事件驱动的,事件与时间密切相关,然而目前的事件描述语言不能表达时间概念.文章引入了一组一元时序算子,将传统的事件表达式扩展到时序空间,并给出了时序事件表达式的时间语义.为实现对于复合时序事件的增量式检测,定义了标志取代Petri网(TRPetrinet),它可表示最近发生事件消耗语义,并在网上计算和传递事件发生时间.文中给出了相关的数据结构和检测算法.实验结果表明时序算子可以有效地实现冲突消解,应用TRPetri网对复合时序事件建模和检测具有描述简洁和速度快的优点.
The rules in active databases are eventdriven. While event has close correlation with time, the present event languages are not rich enough to capture temporal concept. This paper introduces a set of unary temporal event operators that extend classical event expressions to temporal space. Time based semantics for temporal event expressions is presented. A token replacing Petri net (TR Petri net) is defined for the incremental detection of composite temporal events. The extended Petri net can model more recent occurrence semantics of event consumption, calculate and pass timestamps of event occurrence. The internal data structure as well as an event detection algorithm is also given. Experiment results demonstrate that temporal event operators are effective for conflict resolution and modeling and detecting composite temporal event using TR Petri net have the advantages of neat and speed.
出处
《系统工程学报》
CSCD
2003年第3期262-267,共6页
Journal of Systems Engineering
基金
国家自然科学基金资助项目(69673015)
吉林省科技发展计划资助项目(20000111).
