文摘
Bugs prioritization in open source repositories poses as a challenging and complex task, given the significant number of reports and the impact of a wrong bug assignment to the software evolution. Deciding the most suitable bugs in order to be solved can be considered as an optimization problem. Thus, we propose a search-bas ed approach supported by a multi-objective paradigm to tackle this problem, aiming to maximize the resolution of the most important bugs, while minimizing the risk of later resolution of the most severe ones. Furthermore, we propose a strategy to avoid the developer’s effort when choosing a solution from the Pareto Front. Regarding the empirical study, we evaluate the performance of three metaheuristics and investigate the human competitiveness of the approach. Overall, the proposal can be said human competitive in a real-world scenario and the NSGA-II outperformed both MOCell and IBEA in the adopted quality measures.