refactoring

Some factors can change the software and affect the quality, such as the new users' requirements and the need for compatibility with modern techniques. These factors impose a high cost on technical software maintenance. One of the techniques for software quality improvement and maintenance cost reduction is refactoring. The advantage of this method is software behavior preservation. Because the cost of refactoring manually is high, a technique called the hybrid optimization problem has been proposed. The main challenge in refactoring is to propose a technique with high accuracy and less runtime. Hence, in the present work, a refactoring method based on the multi-objective algorithms called RMMOC is proposed to tradeoff between quality and runtime. In this method, a helpful search-based method called UMOCell is used to increase refactoring quality. This method inspires both population-based and local-based search algorithms. Another novelty in this paper is using new metrics for program quality assessment that help to increase accuracy, decrease runtime of refactoring, and find the best refactoring solutions. Because software metrics play a significant role in search-based refactoring approaches, this paper introduces two effective criteria called MPC and refactoring number reduction in addition to previous presented metrics. The results of experiments show that the performance of the proposed method is remarkable and using new metrics is effectiveness.

Ontology inconsistency is one of the most important topics in the field of ontology matching. Until now many matchers are introduced but most of them suffer from inconsistencies. Many of the ontology matching tools have severe problems with respect to the quality of matching results and therefore the results of matching process is not adequate. In this paper, we focus on this topic and present a new method to produce better results from the matching process. The major novelty of this paper is in detecting the inconsistencies in ontologies before starting the matching process. In this phase, many problems caused by ontology diversity are resolved. Besides, some new patterns and inconsistencies in ontologies are detected and then refactoring operations are applied on them. At the end, one of the well-known matchers in OAEI is selected to evaluate our work. Experimental results show that the transformed ontologies are more efficient than original unrepaired ones with respect to the standard evaluation measures.