Metamorphic Relation Automation: State of the Art in Detection, Selection, and Generation Over Two Decades
Year: 2025
Type: preprint
Abstract: Metamorphic testing (MT) has emerged as a promising technique for alleviating the test oracle problem by determining software behavior through metamorphic relations (MRs) without the need for test oracles. The effectiveness of MT crucially relies on identifying appropriate MRs. The traditional process of identifying MRs required extensive domain knowledge and was labor-intensive. To overcome these limitations, various semi-automation and automation techniques have been proposed, such as metamorphic relation composition, metamorphic relation patterns, etc. Therefore, a systematic organization and synthesis of these techniques, covering principles, applications, evaluations, merits and limitations, and future prospects, remains necessary. Through a comprehensive review of 88 publications on metamorphic relation automation, this paper categorizes the field into three subfields: detection, selection, and generation, and systematically analyzes them in terms of technological evolution, evaluation analysis, as well as challenges and future directions. The technological evolution summarizes the classification, applicability, strengths and weaknesses of all relevant technologies in each subfield. The evaluation analysis includes subject programs, MR expression types, common metrics used to assess the effectiveness of these techniques and fault types that may impact the effectiveness of MR automation. Additionally, the challenges and future directions identify potential areas for further research to address ongoing difficulties within MR automation. This comprehensive review aims to provide a thorough understanding of the current landscape and offer insights for researchers and practitioners in the field of MR automation.
Language: English
Subfield: Software
Field: Computer Science
Domain: Physical Sciences
Open Access status: gold
Award: G400648788
OpenAlex ID: W4414457733
Cites:
Cited by:
Related to: