Abstract: Real-time transactions will restart or abort possibly for competing limited system resources in low bandwidth environments
and the unpredictable system workloads in mobile environments.Conventional scheduling algorithms of real-time transactions
cannot adapt this situation，so a novel scheduling algorithm of real-time transactions is proposed based on traversing a directed
acyclic graph which describes the derived relationship of all real-time data objects.Some update transactions with low value
will be discarded to reduce system workload considering the data objects validity in time domain and value domain.The
deadline miss ratio of transactions can be reduced and data freshness can be improved through simulation experiments.

摘要： 移动环境下实时数据库系统负载的不可预测，以及实时事务争夺有限的系统资源经常导致实时事务重启或夭折。传统的实时事务调度算法已不适应，在用有向非循环图表示数据相互间的导出关系的基础上，提出一种基于遍历这种图的实时事务调度算法。结合实时数据对象的时间域和值域有效性，系统适当地丢弃一些低价值的更新事务以减轻系统负载。仿真实验表明：算法一定程度上降低了事务错过截止期比率并提高了数据新鲜度。