Abstract:
ABSTRACT
SELF CORRECTING SCHEDULE
by
Chris McIntosh
Master of Science in Computer Science
California State University, Chico
Summer 2009
Many cutting edge companies find it difficult to accurately predict task duration
and completion of a new project. This paper documents the second edition of the
Self Correcting Schedule (SCS) that helps mitigate the problem by adapting the schedule
over time and relying on real data points instead of the initial raw estimates.
The original Self Correcting Schedule was a prototype developed internally
by the company under study. The concept of self correction with correction factors was
an idea the company came up with. For this project I rewrote the tool into a maintainable
framework and implemented key new features to extend the functionality of the
tool.
The SCS can calculate two project completion dates, the long pole date, and
load balanced date. The definition of each method is:
viii
1. Long pole method: New tasks are assigned to a specific engineer. Therefore the
completion date for the entire project will be when the last engineer finishes all of his or
her tasks.
2. Load Balanced Method: New tasks are assigned to any available engineer until
all tasks in the project are completed. Therefore the project completion date is the date
of the last task being completed.
The SCS will calculate a correction factor based on a formula of estimated
work versus actual work. This factor can then be applied to both the Long pole or Load
balanced dates to yield corrected dates. These dates are more accurate than just adding
the original dates, and when used together can be used to not only track the project but
to detect staffing issues early on in the project.
