An intelligent test management system for optimizing decision making during software testing
Artikel i vetenskaplig tidskrift, 2025

To ensure the proper testing of any software product, it is imperative to cover various functional and non-functional requirements at different testing levels (e.g., unit or integration testing). Ensuring appropriate testing requires making a series of decisions—e.g., assigning features to distinct Continuous Integration (CI) configurations or determining which test specifications to automate. Such decisions are generally made manually and require in-depth domain knowledge. This study introduces, implements, and evaluates ITMOS (Intelligent Test Management Optimization System), an intelligent test management system designed to optimize decision-making during the software testing process. ITMOS efficiently processes new requirements presented in natural language, segregating each requirement into appropriate CI configurations based on predefined quality criteria. Additionally, ITMOS has the capability to suggest a set of test specifications for test automation. The feasibility and potential applicability of the proposed solution were empirically evaluated in an industrial telecommunications project at Ericsson. In this context, ITMOS achieved accurate results for decision-making tasks, exceeding the requirements set by domain experts.

Software testing

Natural language processing

Machine learning

Decision support

Continuous integration

Författare

Albin Lönnfält

Student vid Chalmers

Viktor Tu

Student vid Chalmers

Gregory Gay

Chalmers, Data- och informationsteknik, Interaktionsdesign och Software Engineering

Göteborgs universitet

Animesh Singh

Ericsson AB

Sahar Tahvili

Mälardalens högskola

Ericsson AB

Journal of Systems and Software

0164-1212 (ISSN)

Vol. 219 112202

Ämneskategorier

Programvaruteknik

Datavetenskap (datalogi)

DOI

10.1016/j.jss.2024.112202

Mer information

Senast uppdaterat

2024-10-02