Data-centric Approach to Decision Making in Semiconductor Manufacturing: Best Practices and Future Directions

Abstract—Data-centric approach to decision making has become increasingly crucial in semiconductor manufacturing, revolutionizing the industry's approach to efficiency, quality control, and cost reduction. The integration of advanced analytics, machine learning, and artificial intelligence enables real-time monitoring, predictive maintenance, and adaptive control systems, thereby minimizing downtime, reducing waste, and improving the overall equipment effectiveness. This study explores various types of data collected in semiconductor manufacturing, such as process parameters, equipment sensors, yield data, and quality metrics, and examines the role of advanced analytics techniques in extracting insights from these data. The importance of real-time data processing and analysis for rapid decision-making in semiconductor fabs is highlighted, along with the challenges of data quality, integration, and governance. The study also addresses the use of data visualization tools and techniques to present complex manufacturing data in an easily understandable format for decision-makers. Case studies of successful data-centric approaches in semiconductor manufacturing are examined, showing the benefits and lessons learned. The role of Industry 4.0 and the Internet of Things in enabling more comprehensive data collection and analysis is discussed, as well as the potential of edge computing and fog computing in processing data closer to the source. The integration of supply chain data with manufacturing data for more holistic decision making is explored, and the human factors in data-driven decision making, including the need for training and upskilling of the workforce, are addressed. Finally, the paper concludes with a discussion of future directions, including emerging technologies and trends that may shape data-centric decision-making in semiconductor manufacturing, such as advanced artificial intelligence, cellular networks, quantum computing, digital twins, and focus on sustainability and energy efficiency. Keywords— data-driven decision making, semiconductor manufacturing, advanced analytics, machine learning, artificial intelligence, predictive maintenance, adaptive control systems, overall equipment effectiveness (OEE), equipment sensors, yield data, quality metrics, real-time data processing, data quality and governance, data visualization