Title: Artificial Intelligence Across Disciplines: Research, Innovation, and Intelligent Solutions
Editors: Dr. Subita Bhagat, Dr. A. Balamurugan, Dr. P. Krishna Kumar, and Mrs. S. Nandhini Devi
ISBN: 978-81-69857-83-3
Chapter: 21
DOI: https://doi.org/10.59646/815/21
Author: Y Joanspreetha
Abstract
Artificial intelligence (AI) has fundamentally restructured software engineering by substituting static heuristic processes with rigorous mathematical, probabilistic, and algorithmic optimization frameworks. This chapter develops a unified mathematical foundation for intelligent software engineering across three critical phases of the software development lifecycle: predictive defect classification via regularized logistic formulations, automated test suite minimization utilizing multi-objective combinatorial optimization, and continuous integration runtime scheduling via Markov decision processes. Evaluated across an enterprise-scale software repository comprising over 2.4 million source lines of code and 18,500 continuous integration builds, the integrated framework demonstrates an 88.4% defect classification accuracy (-score: 0.862), compresses test execution suites by 64.2% while preserving 99.1% of mutation fault detection coverage, and curtails regression pipeline build latency by 51.7%. Through formal mathematical models, step-by-step sample calculations, and empirical performance evaluations, this chapter establishes reproducible engineering guidelines for deploying robust, scalable, and provably bounded machine learning architectures within modern mission-critical software ecosystems.
Keywords: Search-Based Software Engineering, Defect Prediction, Test Suite Minimization, Multi-Objective Optimization, Markov Decision Processes, Continuous Integration.