| 摘要: |
| 目的 日志修复是业务流程管理的主要研究内容之一,当业务流程模型中包含资源约束时,以控制流为基础
的流程建模技术和以往的控制流修复方法难以察觉模型中资源视角存在的问题;因此提出了基于一致性检测的资
源感知日志修复方法。 方法 首先使用 Petri 网对流程进行建模,然后将资源视角融入已有 Petri 网中形成流程的资
源感知 Petri 网,建立流程中活动与资源的关联关系;再根据对齐方法分析所产生事件日志的控制流和数据流的偏
差,最后对资源数据异常的事件日志进行分类修复。 结果 以城市轨道全自动驾驶系统的日常运营流程融合故障处
理流程为例验证了方法的可行性,并根据迹的拟合度扩展出复合拟合度,拟合度能够计算控制流视角和资源视角
对模型拟合的综合影响程度;提出的方法与传统局限于控制流对齐的修复方法相比,其复合拟合度更高,日志修复
的结果更好。 结论 通过使用加入资源约束的流程建模技术,并将传统的只针对控制流进行日志修复的方法拓展到
数据流方面,可以使流程中资源视角下可能发生的数据异常不再被忽略,扩大了日志修复的适用范围,有利于解决
日志中资源方面可能存在的问题。 |
| 关键词: 日志修复 一致性检测 资源感知 全自动驾驶系统 流程模型 |
| DOI： |
| 分类号: |
| 基金项目: |
|
| A Resource-aware Log Repair Method Based on Consistency Checking |
|
CHEN Yongqi FANG Na
|
|
School of Mathematics and Big Data Anhui University of Science and Technology Anhui Huainan 232001 China
|
| Abstract: |
| Objective Log repair is one of the main research topics in business process management. When business
process models include resource constraints control-flow-based process modeling techniques and traditional control-flow
repair methods struggle to detect issues from a resource perspective in the model. Therefore a resource-aware log repair
method based on consistency detection was proposed. Methods Firstly the process was modeled using Petri nets and
then the resource perspective was integrated into the existing Petri net to form a resource-aware Petri net of the process.
The association between activities and resources in the process was established. Then deviations between the control flow
and data flow of the generated event logs were analyzed using alignment methods and finally event logs with resource
data anomalies were classified and repaired. Results The feasibility of the method was verified by integrating the daily
operation process of the fully automated urban rail driving system with the fault handling process and the composite fitness
was extended based on trace fitness. This fitness can calculate the comprehensive impact of the control flow perspective
and the resource perspective on model fitting. Compared with traditional repair methods limited to control-flow alignment
the proposed method achieves higher composite fitness and better results in log repair. Conclusion By using process
modeling techniques with added resource constraints and extending traditional log repair methods which focus only on
control flow to the data flow aspect potential data anomalies from a resource perspective in the process are no longer
ignored expanding the scope of log repair and helping to address potential resource-related issues in logs. |
| Key words: log repair consistency detection resource awareness fully automated driving system process modeling |
