MultiRefactor: Automated Refactoring To Improve Software Quality

📝 Abstract

In this paper, a new approach is proposed for automated software maintenance. The tool is able to perform 26 different refactorings. It also contains a large selection of metrics to measure the impact of the refactorings on the software and six different search based optimization algorithms to improve the software. This tool contains both mono-objective and multi-objective search techniques for software improvement and is fully automated. The paper describes the various capabilities of the tool, the unique aspects of it, and also presents some research results from experimentation. The individual metrics are tested across five different codebases to deduce the most effective metrics for general quality improvement. It is found that the metrics that relate to more specific elements of the code are more useful for driving change in the search. The mono-objective genetic algorithm is also tested against the multi-objective algorithm to see how comparable the results gained are with three separate objectives. When comparing the best solutions of each individual objective the multi-objective approach generates suitable improvements in quality in less time, allowing for rapid maintenance cycles.

💡 Analysis

In this paper, a new approach is proposed for automated software maintenance. The tool is able to perform 26 different refactorings. It also contains a large selection of metrics to measure the impact of the refactorings on the software and six different search based optimization algorithms to improve the software. This tool contains both mono-objective and multi-objective search techniques for software improvement and is fully automated. The paper describes the various capabilities of the tool, the unique aspects of it, and also presents some research results from experimentation. The individual metrics are tested across five different codebases to deduce the most effective metrics for general quality improvement. It is found that the metrics that relate to more specific elements of the code are more useful for driving change in the search. The mono-objective genetic algorithm is also tested against the multi-objective algorithm to see how comparable the results gained are with three separate objectives. When comparing the best solutions of each individual objective the multi-objective approach generates suitable improvements in quality in less time, allowing for rapid maintenance cycles.

📄 Content

MultiRefactor: Automated Refactoring To Improve Software Quality Michael Mohan (0000-0002-9944-5821) and Des Greer Queen’s University Belfast, Northern Ireland, UK {mmohan03, des.greer}@qub.ab.uk Abstract. In this paper, a new approach is proposed for automated software mainte- nance. The tool is able to perform 26 different refactorings. It also contains a large selec- tion of metrics to measure the impact of the refactorings on the software and six differ- ent search based optimization algorithms to improve the software. This tool contains both mono-objective and multi-objective search techniques for software improvement and is fully automated. The paper describes the various capabilities of the tool, the unique aspects of it, and also presents some research results from experimentation. The individual metrics are tested across five different codebases to deduce the most effective metrics for general quality improvement. It is found that the metrics that relate to more specific elements of the code are more useful for driving change in the search. The mono-objective genetic algorithm is also tested against the multi-objective algorithm to see how comparable the results gained are with three separate objectives. When compar- ing the best solutions of each individual objective the multi-objective approach generates suitable improvements in quality in less time, allowing for rapid maintenance cycles. Keywords: Search Based Software Engineering • Automated Maintenance • Refactoring Tools • Multi-Objective Optimization • Software Metrics. 1 Introduction Search based optimization has been used extensively in various areas of engineering and in recent years has also been applied to software engineering. Search Based Software Engineering (SBSE) is an area of research that attempts to apply search heuristics to solve complex problems in software development [1]. Software maintenance is one of the more expensive parts of the software development cycle [2]. SBSE applied to maintenance, known as Search Based Software Maintenance (SBSM), is used to assist the manual aspects of maintaining a software project and minimize the time necessary to do so. To aid with this research various tools [3–11] have been used to assist with the refactoring of a software pro- ject. An increasing amount of SBSM research is looking at multi-objective techniques [12–20]. Many multi-objective search algorithms are built with genetic algorithms, as their ability to generate multiple possible solutions is suitable for a multi-objective approach. Instead of focusing on only one property, the multi-objective algorithm will be concerned with a number of different objectives. The MultiRefactor tool uses refactorings to improve Java projects using metric functions to guide the search. Many of the other tools available have a limited selection of refactorings or metrics available to use. The effort has been made to equip the MultiRefactor tool with a large range of available refactorings and metrics to choose from, in order to promote maximum configurability within the tool. MultiRefactor combines the ability to use a multi-objective approach with the more practical ability to improve the source code itself, while checking the semantics of the refactorings being applied so that the changes in the code are valid with respect to the application domain. In order to assess the capabilities of the MultiRefactor approach, a set of experiments have been set up to compare different procedures available within the tool. Experiments have previously been conducted comparing the other metaheuristic searches [21], so the experimentation here focuses on the use of the genetic algorithms in the tool and aims to find out two things. The first aim is to test the available soft- ware metrics within the tool and discover which are more successful. Some metrics may be more useful than others in measuring the changes made by the available refactorings. These will be more helpful when trying to analyze the changes made to a solution and as such, a metric function made from these metrics may assist in creating a more prosperous solution. The second aim is to compare the mono- objective approach with the multi-objective search available and see whether using a multi-objective algorithm to automate maintenance of a software solution is as practical as using a mono-objective algo- rithm. We wish to test whether, in a fully automated solution, a multi-objective algorithm using similar settings can yield comparable results across all the objectives, and whether it is worth the time taken to do so. The following research questions have been formed to address these concerns, along with a cor- responding set of hypotheses and null hypotheses for each factor investigated in RQ2: RQ1: Which set of software metrics have the most variability when used with a mono-objective genetic algorithm to refactor software? RQ2: Does a multi-obje

This content is AI-processed based on ArXiv data.