返回
Hierarchical System Decomposition Using Genetic Algorithm for Future Sustainable Computing
DOI:10.3390/su12062177.png)
摘要
En 中文
A Hierarchical Subsystem Decomposition (HSD) is of great help in understanding large-scale software systems from the software architecture level. However, due to the lack of software architecture management, HSD documentations are often outdated, or they disappear in the course of repeated changes of a software system. Thus, in this paper, we propose a new approach for recovering HSD according to the intended design criteria based on a genetic algorithm to find an optimal solution. Experiments are performed to evaluate the proposed approach using two open source software systems with the 14 fitness functions of the genetic algorithm (GA). The HSDs recovered by our approach have different structural characteristics according to objectives. In the analysis on our GA operators, crossover contributes to a relatively large improvement in the early phase of a search. Mutation renders small-scale improvement in the whole search. Our GA is compared with a Hill-Climbing algorithm (HC) implemented by our GA operators. Although it is still in the primitive stage, our GA leads to higher-quality HSDs than HC. The experimental results indicate that the proposed approach delivers better performance than the existing approach.
Keyword:
software
software maintenance
subsystem decomposition
architecture
genetic algorithm
artificial intelligence
restructuring
AI总结
对已上传原文的论文进行重点信息的提取,主要内容包括:简要概述、研究摘要、背景介绍、关键亮点、图文解析、展望与总结。
期刊
IF:
3.3
论文数:
10.6W
被引数:
28.4W
机构
引用论文
API-based and information-theoretic metrics for measuring the quality of software modularization用于测量软件模块化质量的基于API和信息论的度量
Metrics for measuring the quality of modularization of large-scale object-oriented software衡量大型面向对象软件模块化质量的指标
Improving modular structure of software system using structural and lexical dependency利用结构和词汇依赖改善软件系统的模块化结构

