Doctor of Philosophy
School of Computing and Information Technology
Process Mining techniques have traditionally focused on post-mortem analysis to optimize business processes, leveraging event logs to identify bottlenecks and inefficiencies. These methods aim to optimize process performance by mining process behavior, identifying bottlenecks, and pinpointing other sources of inefficiencies and defects. Complementing this, Process Analytics extends the analytical framework to support the practice of Business Process Management (BPM). This is achieved by leveraging process-related data to extract knowledge, enhance process performance, and facilitate managerial decision-making across the organization. Although considerable attention has recently been devoted to the development of process mining techniques for interpreting process behavior from event logs, the task of leveraging these logs to provide decision support for achieving optimal process outcomes remains an open challenge. Existing techniques have severe shortcomings, especially in the settings of unstructured, knowledge-intensive processes. Moreover, current methods struggle to provide value to organizations in real-world settings where event logs are geographically dispersed, noisy, and incomplete.
The first part of this thesis addresses these challenges by demonstrating how process data can be employed to prescribe tasks in order to achieve optimal outcomes (in both structured and unstructured knowledge-intensive processes where context plays a crucial role). The second part of the thesis contributes to the field of process mining by specifically tackling the challenges associated with understanding process behavior from real-world, geographically dispersed logs that are often noisy and incomplete. To evaluate the effectiveness of the proposed methods, the research methodology employed in this thesis comprises quantitative experiments using real-world data. Data are collected from archived event logs generated by business process management systems and from publicly available datasets, all in strict adherence to ethical guidelines.
Khan, Asjad, A computational framework for Data-Driven Decision Support and Knowledge Augmented Process Analytics, Doctor of Philosophy thesis, School of Computing and Information Technology, University of Wollongong, 2022. https://ro.uow.edu.au/theses1/1735
FoR codes (2008)
0803 COMPUTER SOFTWARE
Unless otherwise indicated, the views expressed in this thesis are those of the author and do not necessarily represent the views of the University of Wollongong.