College of Graduate Studies: Theses & Dissertations
Term of Award
Spring 2026
Degree Name
Master of Science in Mathematics (M.S.)
Document Type and Release Option
Thesis (restricted to Georgia Southern)
Copyright Statement / License for Reuse
This work is licensed under a Creative Commons Attribution 4.0 License.
Department
Department of Mathematical Sciences
Committee Chair
Charles Champ
Committee Member 1
Andrew Sills
Committee Member 2
Ahmed Al-Taweel
Abstract
Classical Shewhart control charts are widely used for monitoring manufacturing processes, but their reliance on single-point signals can delay detection of moderate and sustained shifts. Supplementary runs rules improve sensitivity by capturing nonrandom patterns, yet they are typically applied in a fixed, heuristic manner and are not easily adapted to changing process conditions. This thesis develops a unified framework that integrates runs-rule logic with machine learning to enhance statistical process control while preserving interpretability. Using simulation experiments and data from an electrical component defect screening process, Shewhart, runs-rule augmented, and machine-learning-based monitoring schemes were compared through average run length and detection-delay analysis. The classical chart achieved an in-control ARL near 369, while calibrated runs-rule and machine learning approaches maintained acceptable false-alarm rates and improved detection of moderate shifts. The results demonstrate that runs rules can be formalized as structured features and combined with machine learning to produce interpretable, adaptive, and statistically grounded process-monitoring systems.
Recommended Citation
Garuba, Moshood A., "Integrating Runs-Rule-Based Statistical Process Control with Machine Learning for Adaptive Process Monitoring" (2026). College of Graduate Studies: Theses & Dissertations. 3171.
https://digitalcommons.georgiasouthern.edu/etd/3171
Research Data and Supplementary Material
Yes
