Artificial Intelligence-Enabled Digital Twins for Smart Computing and Real-Time Decision-Making
DOI: https://doi.org/10.5281/zenodo.20590582
Abstract
Artificial intelligence (AI) and digital twin (DT) technologies are increasingly converging to create computational replicas that do more than visualize physical systems. Modern AI-enabled DTs integrate sensor streams, edge-cloud computing, simulation, machine learning, optimization and human feedback to support monitoring, prediction and action in near real time. This review examines how smart computing architectures transform DTs from passive digital shadows into adaptive decision-support systems for computer science and cyber-physical applications. A structured literature synthesis was conducted around peer-reviewed studies on AI in digital twins, real-time analytics, edge intelligence, cybersecurity, smart manufacturing, healthcare, smart cities and energy systems. The paper proposes a layered conceptual architecture, compares AI techniques used across DT functions, and analyses the decision cycle from sensing to actuation. The findings show that supervised learning and deep learning are dominant for predictive maintenance and anomaly detection, while reinforcement learning, optimization, federated learning, knowledge graphs and explainable AI are becoming essential for autonomous, privacy-aware and trustworthy decision-making. However, major barriers remain: data quality, model drift, interoperability, validation, cybersecurity exposure, explainability, latency, regulatory compliance and human accountability. The review argues that future research should move from isolated demonstrators toward standardized, secure, explainable and self-adaptive digital twins that combine physics-based models, data-driven intelligence and governance-by-design. The study contributes a computer-science-oriented synthesis and a future research agenda for AI-DT systems intended to make reliable real-time decisions in complex environments.
How to Cite
Sakshi Sharma, "Artificial Intelligence-Enabled Digital Twins for Smart Computing and Real-Time Decision-Making", Vol. 4, Issue 1, 25-04-2026, pp. 80-95. DOI: https://doi.org/10.5281/zenodo.20590582
