An Automated Approach for Scheduling Bug Fix Tasks
Software projects usually maintain bug repositories where both developers and end users can report and track the resolution of software defects. These defects should be fixed and new versions of the software incorporating the patches that solve them must be released. The project manager must schedule a set of error correction tasks with different priorities in order to minimize the time required to accomplish these tasks and guarantee that the more important issues have been fixed. This problem is recurrent for most software organizations and, given the enormous number of potential schedules, a tool that searches for good schedules may be helpful to project managers. In this work we propose a genetic algorithm using information captured from bug repositories to find near optimal schedules. We evaluated our approach using a subset of the Eclipse bug repository and the results suggested better schedules than the schedule followed by the developers and schedules proposed by a simpler search procedure.