Software Refactoring Scheme Based on NSGA-II Algorithm

As an important software maintenance measure, the key of software refactoring is to improve the unreasonable part of code module, namely code smell. There are many refactoring points with different priorities and severity in the reconstruction activities of large-scale software system. How to determine a suitable reconstruction scheme is an important problem to be solved. In view of the above problems, a multi-objective software reconstruction scheme model is established, which aims at eliminating the number of code smell, the severity of code smell and its importance. Meanwhile, the number of reading code and the difficulty degree are used as the auxiliary decision indicators. NSGA-II algorithm is used to solve the optimal solution set, and the software system JHotDraw is used to verify. Two kinds of scheme sets with higher robustness are obtained considering the fluctuation of severity and importance. Two trade-offs are obtained from the two schemes set. Then, the reconstruction activities are sorted according to the reconstruction cost, which can further improve the purpose, robustness and effectiveness of the code reconstruction activities.