ISRI Seminar Series Announcement

Please join us for our first ISRI Seminar Series, our guest speaker is
Frank Padberg.

Bio:

Frank Padberg received the master's degree in mathematics from the
University of Erlangen, Germany, in 1993, and the Ph.D. degree in
computer science from the University of SaarbrŸcken, Germany, in
1998.  He has since been working with Professor Walter Tichy as a
senior researcher at the Faculty of Informatics, University of
Karlsruhe, Germany. He has been supported by fellowships and research
grants from the German government (Deutsche Forschungsgemeinschaft
DFG).  Before joining the university, he gave seminars within the
industry on networking technology and operating systems.  He has
published papers on different topics in software engineering,
including software process modeling and simulation, effort estimation
and scheduling, software reliability, extreme programming, and
software inspections.

Abstract:

STOCHASTIC SIMULATION AND OPTIMIZATION OF SOFTWARE PROJECT SCHEDULES

The time and effort needed to complete a software development project
are hard to estimate in practice.  In addition to the uncertainty
about the duration of the individual development steps, software
projects often suffer from unexpected rework and delays. Changes and
rework in some part of the software usually lead to additional rework
somewhere else ("ripple effects"). As a result, software project
estimation and - even more so, software project scheduling - remain
notoriously difficult tasks for the managers.

I'm addressing the estimation and scheduling problem for software
projects using a stochastic scheduling model. The model represents an
explanatory theory of how unexpected rework propagates through a
software project and leads to delays and increased cost. Given certain
statistical data about past projects and high-level design data about
the software being developed, the model allows a manager to compute
probability distributions for the project completion time and
cost. These distributions also quantify the risk of not meeting a
given deadline or budget.

In the talk, I'll outline my scheduling model and show how discrete
process simulation helps to understand the dynamics of ripple effects
in a project. I'll also show how simulation can be applied to evaluate
and compare different scheduling policies for a given project. At the
end of the talk, I want to sketch how to compute optimal scheduling
policies in the model and present first computational results for some
sample projects.


Where: Wean Hall 4623
When: 9/9/04
Time: 12:00 - 1:30

LUNCH PROVIDED