透過您的圖書館登入
IP:18.116.63.236
  • 學位論文

運用工作流程探勘於企業工作流程之異常偵測

Using Workflow Mining on Outlier Detection of Business Workflow

指導教授 : 賴錦慧

摘要


隨著企業營運所涉及的流程愈趨繁雜,為了要讓企業流程能為公司帶來更好的工作效率,本研究提出了各種檢視工作流程異常的方法。首先,本研究使用工作流程探勘方法,從工作流程執行記錄發掘工作流程模型,並找出未發生與不應發生的異常工作樣式。 在異常偵測部分,本研究則提出了兩個方法:在單一工作流程中,以活動的執行時間與等待時間,結合統計的平均值、標準差和發生異常的頻率,找出異常程度較高的任務;針對所有工作流程的部分,以流程的總時間為基礎,計算流程的平均任務時間、標準差以及發生異常的頻率,找出總流程時間異常程度較高的工作流程。企業可透過上述方法所產生的結果,進而探討造成異常的原因。 本研究並以個案公司的系統記錄進行實測,有效地幫助個案公司迅速找出問題點。因此,透過本研究所提出之方法,企業不論是針對單一工作流程或是企業整體流程,皆能協助管理者找出工作流程瓶頸,進而提出正確的改善方案。

並列摘要


Because the workflows in enterprise operations are very complicated, it is difficult to automatically identify the anomalies in workflow. To improve the efficiency of workflow, this work proposes the outlier detection methods in business processes for automatically analyzing the anomalous tasks and workflow patterns. First, workflow mining method is used to identify general workflow models from workflow log. Then, two algorithms of outlier detection for the workflow models are proposed. The one can detect high anomalous tasks in a workflow by considering task execution time, waiting time, and some statistics metric (i.e., average time, standard deviation and anomaly frequency). The other can detect anomalous workflow patterns from all of workflow models according to the average execution time of a task in all workflow, standard deviation and anomaly frequency. Based on a real scenario in an enterprise, the experimental results show that the proposed methods can detect and discover the outliers of workflow accurately. Therefore, this work may support the managers of an enterprise to quickly find out the problems in business processes. It also can help them propose some appropriate solutions for improving a single workflow or business processes.

參考文獻


鍾睿軒(2007)。探勘工作流程之改變樣式。國立中央大學企業管理研究所,桃園縣。
黃文範(2006)。由執行記錄中探勘具備活動期間之工作流程模型。國立政治大學資訊科學學系,台北市。
van der Aalst, W. M. P., Weijters, A. J. M. M., & Maruster, L. (2002). Workflow mining: Which processes can be rediscovered? BETA Working Paper Series, WP 74: Eindhoven University of Technology.
Allen, R. (2001). Workflow: an introduction. Workflow handbook, 15-38.
Cook, J. E., & Wolf, A. L. (1998). Discovering models of software processes from event-based data. ACM Transactions on Software Engineering and Methodology (TOSEM), 7(3), 215-249.

延伸閱讀