Integrating Machine Learning Techniques across Project Management: Enhancing Decision Making and Risk Mitigation

Abstract

The traditional forms of project management (PM) are not able to cope with the changing conditions as project environments become more complex, marked by increased uncertainty, distributed work teams, dynamic stakeholder requirements and unstable resource environments. Machine learning (ML) provides the ability to perform data-driven forecasting, anomaly detection, optimisation of resources, and prediction of a scenario with powerful capabilities. This review paper is a synthesis of existing studies on the use of ML in project management with respect to two fundamental advantages, including improved decision making and risk mitigation. To begin with, it investigates the use of ML methods (supervised, unsupervised, reinforcement) at PM phases (initiation, planning, execution, monitoring, closing). It then discusses the implications in risk management, which are early risk identification, adaptive risk response, real time monitoring, and predictive risk scoring. Next, the article pinpoints some of the enablers (data availability, integration with PM systems, organisational culture) and obstacles (data quality, model interpretability, ethical/trust issues). Lastly, it suggests an idealized paradigm of applying ML to PM practices and future research areas like human in the loop ML, explainable ML and longitudinal impact studies.