Industrial Validation of a Neural Network Model Using the Novel MixTCP Tool

Arnold Szederjesi-Dragomir, Radu Găceanu, Andreea Vescan

2024

Abstract

Test Case Prioritization (TCP) is crucial in the fast-paced world of software development to speed up and optimize testing procedures, particularly in Continuous Integration (CI) setups. This paper aims to first validate a state-of-the-art neural network model to TCP in CI environments, by applying it into a real-world industrial context, and second to propose MixTCP, a tool that integrates the neural network model and significantly enhances the regression testing experience from the software developer perspective. MixTCP is implemented in the Elixir programming language and employs the NEUTRON model, a state-of-the-art approach that uses neural networks to intelligently prioritize test cases, effectively improving fault detection and reducing testing time. The tool is composed of loosely coupled components (Mix TCP task, TCP Server, and NEUTRON model), thus enabling the integration of other Test Case Prioritization solutions too. The results show that MixTCP has the potential to be a valuable asset to modern software development methods, offering software engineers a more efficient, a more user-friendly, and an overall easier to integrate TCP approach.

Paper Citation

in Harvard Style

Szederjesi-Dragomir A., Găceanu R. and Vescan A. (2024). Industrial Validation of a Neural Network Model Using the Novel MixTCP Tool. In Proceedings of the 19th International Conference on Evaluation of Novel Approaches to Software Engineering - Volume 1: ENASE; ISBN 978-989-758-696-5, SciTePress, pages 110-119. DOI: 10.5220/0012631100003687

in Bibtex Style

@conference{enase24,
author={Arnold Szederjesi-Dragomir and Radu Găceanu and Andreea Vescan},
title={Industrial Validation of a Neural Network Model Using the Novel MixTCP Tool},
booktitle={Proceedings of the 19th International Conference on Evaluation of Novel Approaches to Software Engineering - Volume 1: ENASE},
year={2024},
pages={110-119},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0012631100003687},
isbn={978-989-758-696-5},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 19th International Conference on Evaluation of Novel Approaches to Software Engineering - Volume 1: ENASE
TI - Industrial Validation of a Neural Network Model Using the Novel MixTCP Tool
SN - 978-989-758-696-5
AU - Szederjesi-Dragomir A.
AU - Găceanu R.
AU - Vescan A.
PY - 2024
SP - 110
EP - 119
DO - 10.5220/0012631100003687
PB - SciTePress