Mutation Testing for Large-Scale Software Systems: Improving Test Suite Effectiveness Using Intelligent Fault Injection

Abstract

ABSTRACT Mutation testing is widely recognized as one of the most effective techniques for evaluating the fault-detection capability of software test suites. By injecting artificial faults (mutants) into source code and observing test behavior, software engineers gain direct insights into test coverage gaps. However, applying conventional mutation testing to enterprise, large-scale software systems presents severe computational bottlenecks due to massive mutant volumes, high execution overhead, and persistent equivalent mutants. This review paper systematically examines recent advances in intelligent fault injection techniques designed to mitigate these constraints. We investigate machine learning-driven mutant selection, static and dynamic code embedding abstractions, predictive mutation scoring, evolutionary test optimization, and parallel execution frameworks. Synthesizing empirical findings across enterprise repositories and open-source benchmarks, this paper evaluates how intelligent mutant prioritization optimizes test suite effectiveness while drastically curtailing execution costs. Finally, we outline research gaps, key tool limitations, and actionable directions for future research in AI-augmented software quality assurance.

KEYWORDS: Mutation Testing; Intelligent Fault Injection; Test Suite Effectiveness; Machine Learning in Software Engineering; Equivalent Mutants; Software Reliability; Large-Scale Systems.