Artificial Intelligence in Project Scheduling Management: A Systematic Literature Review

Abstract

The integration of artificial intelligence into project management practices represents one of the most significant technological transformations in contemporary scheduling methodologies. As projects grow increasingly complex and face greater environmental uncertainty, traditional scheduling approaches often prove inadequate. This systematic literature review examines how AI applications are reshaping project scheduling management, investigating which techniques are being employed and their measurable impact on scheduling effectiveness across diverse project environments. We implemented a rigorous systematic review methodology following Fink's seven-step model, incorporating an innovative triangulated screening approach that combined manual evaluation with AI-assisted validation techniques. Through comprehensive quality assessment and methodological evaluation of initially identified 195 articles, 66 studies published between 2015-2024 were ultimately selected for in-depth analysis. Our analysis identified 10 distinct AI fields, Evolutionary Algorithms, Neural Networks, Swarm Intelligence, Optimization and Metaheuristics, Classical Machine Learning Models, Fuzzy Systems, Reinforcement Learning, Agent-Based Systems, Generative Pre-trained Transformers, and Knowledge Representation and Reasoning, comprising 70 unique techniques applied across 9 core project scheduling functions, including Resource Management, Constraint-Based Scheduling, Time Estimation and Duration Prediction, Task Scheduling and Sequencing, Multi-Objective Scheduling, Uncertainty Management and Stochastic Scheduling, Dynamic and Real-Time Scheduling, Critical Path Identification, and Schedule Optimization. Evolutionary Algorithms, particularly Genetic Algorithms, demonstrated pronounced dominance, while hybrid-method implementations consistently outperformed single-method approaches across key performance metrics. Moreover, the construction industry showed a substantial majority (62%) of applications, raising important questions about cross-sector generalizability of current research findings. This review contributes theoretically through a novel taxonomic framework for classifying AI applications in project scheduling contexts and identification of divergent adoption trajectories between established and emerging methodologies. From a practical perspective, our implementation maturity assessment revealed concerning gaps, with only one-third of studies reaching experimental validation with real-world data. This maturity limitation suggests practitioners should approach technique selection contextually rather than implementing trending approaches without consideration of their specific scheduling challenges and organizational readiness. Looking forward, the field would benefit substantially from development of standardized frameworks that effectively transition theoretical models into operational implementations and from comparative analyses across diverse sectors, enhancing understanding of industry-specific scheduling constraints and their influence on optimal AI technique selection.