Multi-objective optimization and integrated indicator-driven two-stage project recommendation in time-dependent software ecosystem

Abstract

Context:Time-dependent software ecosystem is a complex system, where there are many projects and developers. Recommending projects to developers in a time-dependent software ecosystem can improve their quality and development speeds. However, the time-dependence of projects and developers results in an increased difficulty of project recommendation. Objective:To better recommend projects to developers in a time-dependent software ecosystem, we propose a method of multi-objective optimization and integrated indicator-driven two-stage project recommendation, which is fulfilled according to the change of developer communities and their projects. Method:According to the change of developer communities and their projects, a method of multi-objective optimization and integrated indicator-driven two-stage project recommendation is fulfilled. In the first stage, a constrained multi-objective optimization model for project recommendation to developer communities is established, and an improved NSGA-II algorithm is adopted to solve this model, with the purpose of obtaining the recommended projects to a developer community. For the second stage, an integrated indicator for project recommendation to developers is built to determine the developers of a project. Results:The proposed method is applied to project recommendation for nine time-dependent software ecosystems in GitHub, and compared with six state-of-the-art ones. The experimental results show that our method has significant advantages in recommendation accuracy and efficiency. Conclusion:According to the experimental results, we conclude that the proposed method can timely and accurately recommend projects to developers in a time-dependent software ecosystem, which reduces the difficulty of solving the problem of project recommendation.