Robust scheduling for large projects
Badr, Salah M.
Berzines, Valdis Andris
MetadataShow full item record
We have developed a heuristic scheduling method suitable for automatically scheduling tasks in a software development effort and assigning them to designers. Our experimental evaluations of the algorithm show that it is highly effective at finding feasible schedules when they exist. A modification of the algorithm can suggest nearly minimal adjustments to the deadlines in cases where no feasible schedule exists. This is useful because it provides guidance to the project manager for formulating a proper response when a project gets late and all of the planned tasks cannot be completed with their deadlines. The algorithms are fast enough to support constant rescheduling as circumstances change, for most projects of practical size.
Approved for public release; distribution is unlimited.