home · mobile · calendar · defenses · 1998-1999 · 

Thesis Defense - Keddara

Dynamic Evolution within Workflow Systems
Karim Keddara
Computer Science PhD Candidate

Dynamic change is a large and pervasive problem which surfaces within organizational workflows as well as within software engineering, manufacturing and numerous other domains. Procedural change, performed in an ad hoc manner, can result in surprising negative effects (inefficiencies, inconsistencies, breakdowns ...).

This thesis is concerned with dynamic change to procedures in the context of workflow systems. How can we make workflow systems safely more flexible and open? We believe that part of the problem may be solved using a complete approach that supports the specification, the analysis, the synthesis and the coordination of change. Our work combines all the aforementioned change facets in a framework based on theoretical and pragmatic grounds.

We introduce ML-DEWS, a Modeling Language to Support Dynamic Evolution within Workflow Systems. It is designed specifically with features for the complete and the unambiguous description of procedural change. ML-DEWS is based upon the observation that a process change is a process, and therefore may be modeled, enacted, analyzed, coordinated and monitored as any other process.

We also examine the issue of change analysis in a more formal setting using Timed Flow Nets, a class of Timed Petri Nets. Procedure change is then viewed as a net replacement whereby an old net (representing the old procedure) is transformed into a new net (representing the new procedure). Dynamic change means that the transformation occurs within marked flow nets. Token migration from the old to the new net is facilitated through flow jumpers. Based on this model, various correctness criteria are investigated, e.g. soundness, trace approximation, min-max times, etc.

Committee: Clarence (Skip) Ellis, Professor (Chair)
Andrzej Ehrenfeucht, Professor
Gary Nutt, Professor
Michael Lightner, Department of Electrical and Computer Engineering
Jacques Wainer, Instituto de Computação, Universidade Estadual de Campinas
Department of Computer Science
University of Colorado Boulder
Boulder, CO 80309-0430 USA
May 5, 2012 (14:20)