Document Type : Research Paper
Authors
1
Department of Industrial Engineering, Faculty of Industrial Engineering and Management, Malek Ashtar University of Technology, Tehran, Iran
2
Department of Public Administration, Faculty of Management and Accounting, Allameh Tabataba'i University, Tehran, Iran
10.22054/jims.2026.94136.3051
Abstract
The integration of artificial intelligence (AI) and machine learning (ML) with statistical process control (SPC) offers significant opportunities to enhance process monitoring, pattern recognition, and anomaly detection. However, existing literature remains fragmented across disparate statistical, algorithmic, and operational dimensions. This study systematically reviews research published between 1990 and April 2025 on integrating AI/ML with univariate and multivariate control charts. Following PRISMA guidelines, searches across Scopus and Web of Science (supplemented and verified via Consensus) yielded 134 initial records, resulting in 40 selected studies. Independent screening demonstrated strong inter-rater reliability (85% agreement, Cohen’s kappa = 0.78), followed by a five-criterion quality assessment. The synthesis reveals that AI integration spans four progressive levels: algorithmic integration, data and analytical integration, intelligent monitoring, and adaptive control/decision-making. While deep-learning architectures (notably CNNs, LSTMs, and hybrid models) achieve robust pattern recognition, practical deployment depends heavily on data quality, computational infrastructure, interpretability, and organizational readiness. To bridge theoretical models and operational practice, this study proposes a four-layer conceptual framework comprising the data, AI/ML, control chart, and implementation layers. The framework highlights AI as a multifaceted complement across the SPC lifecycle, providing a structured roadmap toward intelligent, adaptive monitoring.
Keywords
Subjects