Downloads: 0
Study Papers | Computer Science and Engineering | Volume 15 Issue 3, March 2026 | Pages: 958 - 965 | India
Mutation Testing with Machine Learning Approaches for Fuzzing, Smart Contracts, and Automated Repair: A Critical Survey
Abstract: This study critically examines mutation testing enhanced by machine learning across fuzzing systems, browser-based applications, integer overflow detection in smart contracts, and heuristic-based automated program repair. The analysis synthesizes recent literature to identify methodological trends, strengths, and unresolved limitations. Comparative evaluation indicates that reinforcement learning improves input scheduling and mutation efficiency but often lacks semantic understanding and computational scalability. Domain-specific mutation strategies demonstrate effectiveness in vulnerability detection yet remain limited in the process of generalization. The study highlights the need for intelligent mutation frameworks that integrate semantic modelling and cost-aware execution strategies. The findings guide the development of scalable, context-aware mutation testing methodologies.
Keywords: Mutation testing, fuzzing, reinforcement learning, computational scalability, browser testing, vulnerability detection
How to Cite?: Sobhan Kumar Bedajna, Sudipta Roy, "Mutation Testing with Machine Learning Approaches for Fuzzing, Smart Contracts, and Automated Repair: A Critical Survey", Volume 15 Issue 3, March 2026, International Journal of Science and Research (IJSR), Pages: 958-965, https://www.ijsr.net/getabstract.php?paperid=SR26313164836, DOI: https://dx.dx.doi.org/10.21275/SR26313164836