Developer Load Normalization Using Iterative Kuhn-Munkres Algorithm: An Optimization Triaging Approach

TL;DR

This paper presents an automated bug triage approach that optimizes developer workload and bug fixing time by applying an iterative Kuhn-Munkres algorithm for load normalization, outperforming existing systems.

Contribution

It introduces a novel load normalization method for bug triage using an iterative Kuhn-Munkres algorithm, addressing workload imbalance and improving bug fixing efficiency.

Findings

Outperforms existing systems in bug fixing time

Effectively normalizes developer workload

Reduces unaddressed bug reports

Abstract

Bug triage can be defined as the process of assigning a developer to a bug report. The duty of the bug triage team is to study the developers profiles well in order to make an appropriate match between the developers and the incoming bug reports. Thus, this process is a vital step in issue management system. In fact, the number of bug reports submitted every day is gradually increasing which affects the developer workload. Thus, the triage team should consider this factor in distributing the bugs and because of the manual approach, many developers are burden. In particular, triaging bug reports without considering the workload does not only affect the developers workload but also leads to an increase in the number of unaddressed bug reports. As a result, the fixing time of the reported bugs will relatively increase. Unlike other researchers who focus on automating the bug triage and ignoring the developer workload, in this work, we handle the triaging process from a different perspective. The proposed approach focuses on how to optimize the bug fixing time by normalizing the developer load in an automating system. To evaluate our work, we use 26,317 bug reports from different bug repositories. Results shows that our work outperforms other systems in terms of optimizing the bug total fixing time and normalizing developer load.