计算机集成制造系统 ›› 2021, Vol. 27 ›› Issue (9): 2647-2660.DOI: 10.13196/j.cims.2021.09.017

• 当期目次 • 上一篇    下一篇

不完备事件日志下的业务系统变化检测及模型修复方法

方欢,郑雪文,王吴松   

  1. 安徽理工大学数学与大数据学院
  • 出版日期:2021-09-30 发布日期:2021-09-30
  • 基金资助:
    国家自然科学基金资助项目(61572035,61902002);安徽省自然科学基金资助项目(1608085QF149);安徽省高校优秀青年人才基金资助项目(gxyqZD2018038);安徽省博士后基金资助项目(2018B288)。

Change detection and model repair method of business system under incomplete event logs

  • Online:2021-09-30 Published:2021-09-30
  • Supported by:
    Project supported by the National Natural Science Foundation,China(No.61572035,61902002),the Natural Science Foundation of Anhui Province,China(No.1608085QF149),the Excellent Youth Talents Foundation of Anhui Provincial Universities,China(No.gxyqZD2018038),and the Postdoctoral Foundation of Anhui Province,China(No.2018B288).

摘要: 业务系统在演进过程中不可避免地引入变化因素,如何将演进中的变化因素对系统模型进行修复具有实际意义。现有模型修复主要针对原始参考模型和完备日志进行,存在原始参考模型未知和不完备日志的局限性。因此,提出基于不完备日志的变化检测方法,在系统原始模型未知的情形下,利用活动的比较矩阵(行为轮廓矩阵)描述事件日志的Delete、Insert和Move变化操作,从而发现原始事件日志同系统实际日志的局部变化偏差域,根据日志偏差域对原始日志发现的模型进行修复,提出7个实现模型修复的算法。最后,通过ProM工具对所提方法进行了仿真和实验,实验结果表明了方法的有效性和正确性。

关键词: 不完备日志, 行为轮廓, 变化检测, 模型修复, ProM仿真

Abstract: It is inevitable to introduce changing factors in the process of business system evolution,and how to repair the system model by changing factors in the evolution is of practical significance.The existing model repair mainly focuses on the original reference model and complete log,there are limitations of the original reference model unknown and incomplete log.Therefore,a change detection method based on incomplete logs was proposed,and the event log of Delete,Insert and Move operations were described using activity comparison matrix (behavior contour matrix) in the case of the unknown original model system to find the original event log with the local actual log system change bias field.According to the log deviation domain,the model found in the original log was repaired,and seven algorithms for model repair were proposed.The proposed method was simulated and experimented with ProM tool,and the experimental results showed its validity and correctness.

Key words: incomplete log, behavior profile, change detection, model repair, ProM simulation

中图分类号: