home · mobile · calendar · defenses · 2007-2008 · 

Thesis Defense - Rembert

Automatic Discovery of Workflow Models
Aubrey Rembert
Computer Science PhD Candidate
4/30/2008
2:00pm-4:00pm

Workflow systems are model-driven software systems designed to (semi-) automate the coordination of resources, activities, and people within an organization. A workflow system can be a tremendous aid to an organization because it attempts to ensure that the right agent (software, or human) gets the right resource at the right time during the execution of a business process. But, how does a workflow system know the particular business processes, agents, and resources of a particular organization? The answer is business process analysis (BPA) and workflow modeling. The result of BPA and workflow modeling is a workflow model that captures the business processes, agents, and resources of an organization in a rigorous and formal language that can be interpreted by computer software. However, for most organizations, BPA and workflow modeling are time-consuming, error-prone, and expensive tasks that usually result in workflow models that are inaccurate and incomplete.

To mitigate some of the problems traditionally associated with BPA and workflow modeling, many forward-thinking organizations are beginning to record, manually or automatically, the data associated with the executions of their business processes in workflow execution logs that are analyzed during BPA and workflow modeling. These logs, however, are becoming too large to be analyzed manually. This thesis is concerned with workflow discovery, the task of automatically constructing workflow models from workflow execution logs. Specifically, we develop a framework for workflow discovery. This framework considers a business process to be multidimensional and admit observation from multiple perspectives. Using this framework, we've developed algorithms to discover workflow models from the behavioral and informational perspectives of business processes.

Committee: Clarence (Skip) Ellis, Professor (Chair)
Kwang-Hoon Kim, Kyonggi University
Clayton Lewis, Professor
Kenneth Anderson, Associate Professor
Roger (Buzz) King, Professor
Department of Computer Science
University of Colorado Boulder
Boulder, CO 80309-0430 USA
webmaster@cs.colorado.edu
www.cs.colorado.edu
May 5, 2012 (14:20)
XHTML 1.0/CSS2
©2012