Back to Search Start Over

Developer load balancing bug triage: Developed load balance.

Authors :
Yadav, Asmita
Baljon, Mohammed
Mishra, Shailendra
Singh, Sandeep Kumar
Saxena, Sharad
Sharma, Sunil Kumar
Source :
Expert Systems. Jun2024, Vol. 41 Issue 6, p1-25. 25p.
Publication Year :
2024

Abstract

During the software development process, numerous bugs are reported daily in the software bug repositories. Bug triage assigned these bugs to the most relevant and expert developer for resolution. Moreover, assigning bugs to an incompetent or an over‐engaged developer causes repeated reassignment to other developers until it is resolved. This problem can be solved by devising a triage process that assigns bugs to not only expert developers but also to those who are either under‐engaged or reasonably engaged but are not over‐engaged in work. This paper has designed and implemented work engagement sensitive bug triage that resolves the issue of assigning bugs to developers considering their due work engagement, expertise as well as the current state of activity. For this purpose, a developer profile is built by using metrics to generate three types of scores: technical skill, work engagement and work experience. Metadata features like developer‐name, email, developer‐work‐experience in bug resolution, last and present‐work‐activity, timestamp, component and priority are used for it. A multi‐criteria‐based Henry–Garret technique is used to generate a single ranked list of developers from three ranked lists. The performance of the proposed approach is evaluated on four large OOS projects: Mozilla, Eclipse, Netbeans, and Open Office covering 895,439 bug reports accumulated for 33 years of development. The overall system accuracy of the proposed triage using all four datasets is 91.96 ± 0.05% which is 5.87% better than previously published work. The proposed method is achieved up to 90.30 ± 0.05 and 97.1 ± 0.05 MRR and accuracy of reassignment respectively that indicates how significantly it re‐assigns the bug to the relevant developers. The results demonstrate improvement in the accuracy of software‐bugs triaging as well as a reduction in the probability of bug‐tossing. [ABSTRACT FROM AUTHOR]

Details

Language :
English
ISSN :
02664720
Volume :
41
Issue :
6
Database :
Academic Search Index
Journal :
Expert Systems
Publication Type :
Academic Journal
Accession number :
176989629
Full Text :
https://doi.org/10.1111/exsy.13006