Code-Smells Detection as a Bi-Level Problem

Code-Smells represent design situations that can affect the maintenance and evolution of software. They make a system difficult to evolve. Code-smells are detected, in general, using quality metrics that represent some symptoms. However, the selection of suitable quality metrics is challenging due to the absence of consensus to identify some code-smells based on a set of symptoms and also the high calibration effort to determine manually the thresholds value for each metric. In this paper, we propose, for the first time, to consider the generation of code-smells detection rules as a bi-level optimization problem. Bi-level optimization problems are a new class of challenging optimization problems, which contain two levels of optimization tasks. In these problems, only the optimal solutions to the lower level problem become possible feasible candidates to the upper level problem. In this sense, the code-smell detection problem is truly a bi-level optimization problem, but due to lack of suitable solution techniques, it was attempted to solve using a single-level optimization problem in the past. In our adaptation here, the upper level problem generates a set of detection rules, combination of quality metrics, which maximizes the coverage of a base of code-smell examples and artificial code-smells generated by the lower level. The lower level maximizes the number of generated artificial code-smells that cannot be detected by the rules produced by the upper level. The main advantage of our bi-level formulation is that the generation of detection rules is not limited to some code-smell examples identified manually by developers which are difficult to collect but it allows the prediction of new code-smells behavior that are different from those in the base of examples. The statistical analysis of our experiments over 31 runs on 9 open source systems shows that 7 types of code-smell were detected with an average of more than 85% in terms of precision and recall. The results confirm the outperformance of our bi-level proposal compared to state-of-art code-smells detection techniques.