Abstract: Simulation models are useful both as training tools and as decision
support systems. The time compression, 'what if' and graphical representation
capabilities of recent simulation tools such as Extend and Stella give software
engineering simulation models an edge over other techniques. While these
simulation models have been effective in illustrating hypothetical scenarios,
questions remain about how well they can be applied to actual software development
situations. The hypothesis of this research is that a structured approach to the
use of simulation models of the software process will provide valuable insights
regarding alternative software practices.
This can aid in software process improvement. This talk will compare and contrast
continuous simulation models and discrete simulation models. The examination of
discrete models focuses on work done at GTE Labs using Petri Nets to model review
processes as well as software life cycle models. The continuous models are based
on extensive work by Tarek Abdel-Hamid using the system dynamics methodology to
study software process improvement. The talk concludes by posing a research agenda
regarding future work in software process simulation.
Background
Bio: John Trimble is Assistant Professor of Systems and Computer Science at
Howard University. He received his Ph.D. from the Georgia Institute of Technology in 1992. His research has focused on systems analysis, simulation
and software engineering. Trimble has extensive experience in software engineering
and software quality assurance, as well as expert systems, knowledge acquisition
techniques, and system dynamics. He has over 12 years of experience in industry,
and has been the chair of the Computer and Information Science Department at Morris Brown College
in Atlanta.
