A scheduling-driven approach to efficiently assign bug fixing tasks to developers

Abstract The efficient assignment of bug fixing tasks to software developers is of major importance in software maintenance and evolution. When those tasks are not efficiently assigned to developers, the software project might confront extra costs and delays. In this paper, we propose a strategy that minimizes the time and the cost in bug fixing by finding the best feasible developer arrangement to handle bug fixing requests. We enhance therefore a state-of-the-art solution that uses an evolutionary bi-objective algorithm by involving a scheduling-driven approach that explores more parts of the search space. Scheduling is the process of evaluating all possible orders that developers can follow to fix the bugs they have been assigned. Through an empirical study we analyze the performance of the scheduling-driven approach and compare it to state of the art solutions. A non-parametric statistical test with four quality indicator metrics is used to assure its superiority. The experiments using two case-studies (JDT and Platform) showed that the scheduling-driven approach is superior to the state of the art approach in 71% and 74% of cases, respectively. Thus, our approach offers superior performance by assigning more conveniently bug fixing tasks to developers, while still avoiding to overload developers.