一句话：提出 AndroFL 工具，用进化算法生成测试套件并结合统计故障定位，显著降低 Android 应用调试成本。
本文介绍工具 AndroFL，为 Android 应用提供基于进化算法的测试套件生成基础设施，并在其后端接入统计故障定位模块以诊断缺陷。其进化测试生成器支持可配置的适应度函数，例如覆盖率以及 Ulysis 等可诊断性度量；统计故障定位引擎支持 Ochiai、Tarantula、Barinel 等常用指标，并允许用户添加自定义定位指标。作者在来自 F-Droid 的 20 个开源应用上评估 AndroFL，实验表明其相较随机测试显著提升效率：中位 EXAM 分数下降 74%，开发者定位故障的速度约提升 4 倍；在最坏排名情形下，AndroFL 在前 5 和前 10 排名列表中分别比随机测试多定位 25% 和 50% 的故障。
We present our tool, AndroFL, that provides an infrastructure for an evolutionary algorithm-based test-suite generation backed by a statistical fault localization module for diagnosing faults. AndroFL’s evolutionary test-generator supports configurable fitness functions (e.g., coverage, diagnosability metrics like Ulysis). The statistical fault localization engine supports popular metrics like Ochiai, Tarantula and Barinel, and allows adding custom fault localization metrics. We evaluated AndroFL on 20 open-sourced apps from F-Droid, and demonstrates significant efficiency gains: it reduces debugging effort by 74% (median EXAM score) compared to random testing—enabling developers to pinpoint faults ≈ 4× faster. Furthermore, AndroFL localizes 25% and 50% more faults compared to random testing in the top-5 and top-10 ranked list in worst case ranking scenario.
DOI 原文 ·
@inproceedings{paperbot9388,
title = {AndroFL: Evolutionary-Driven Fault Localization for Android Apps},
author = {Vishal Singh and Ravi Shankar Das and Prajwal H G and Subhajit Roy},
booktitle = {ASE},
year = {2025},
doi = {10.1109/ase63991.2025.00353}
}