A New Framework for Software Architecture Generation and Fault Rectification in Software Engineering

摘要

In recent days, the size and complexity of the software system is increased, so designing and specifying the overall software system emerges as a new kind of problem. Most of the modern platforms help to simplify the task of building application, but it does not helps to replace the need for generating the software architecture carefully based on the specific requirements. Though, the achievements of software architecture are challenging, still some issues remain. The complexity, high cost and design erosions are some of the major issues of software architecture. In order to overcome these issues, a new technique, namely, Software Architecture Generation Model - Fault Rectification (SAGM - FR) is proposed in this paper. The proposed system includes the following stages, data preprocessing, normalization, similarity matching, grouping, software fault identification and architecture design. At first, the input data is preprocessed to clean the data by smoothing the noisy data, identifying the outliers and resolving the inconsistencies. Then, the preprocessed data will be normalized to eliminate the redundant data and ensure the data dependency. Here, the similarity matching is performed to construct a logical semantic similarity function after obtaining the semantic relations. Hence, the faults present in the system is rectified and the software architecture is generated. Finally, the evolutionary process is carried out to evaluate the performance of the proposed system. The experimental analysis shows the results of the proposed software architecture generation system in terms of memory utilization, execution time, software dependency, error rate, flexibility and reliability.