Enhanced Adaptive Thresholding LASSO Control Chart Using Resubmission Method for Detecting Covariance Matrix Deviations in High-Dimensional Processes

Fathian, Anahita; Maleki, Mohammad Reza; Eghbali, Hossein

doi:10.22070/jqepo.2026.20792.1307

Enhanced Adaptive Thresholding LASSO Control Chart Using Resubmission Method for Detecting Covariance Matrix Deviations in High-Dimensional Processes

Document Type : CFP- Quality Engineering Techniques in Production and Service Systems

Authors

Anahita Fathian ¹

Mohammad Reza Maleki ²

Hossein Eghbali ¹

¹ Department of Industrial Engineering, University of Eyvanekey, Semnan, Iran.

² Industrial Engineering Group, Golpayegan College of Engineering, Isfahan University of Technology

10.22070/jqepo.2026.20792.1307

Abstract

With the increasing complexity of modern manufacturing systems and the growing dimensionality of quality data, the need for effective tools to track covariance matrix shifts in high-dimensional processes has become more urgent. This study proposes an enhanced adaptive thresholding LASSO-based control chart that utilizes a resubmission technique to enable faster detection of covariance matrix shifts in high-dimensional processes. The performance of the proposed chart is compared to the classical adaptive thresholding LASSO control chart through Monte Carlo simulations. To this end, seven out-of-control scenarios are defined for the covariance matrix, including two diagonal cases, two purely off-diagonal cases, and three mixed diagonal/off-diagonal cases. The detection power of both the proposed and existing control charts is evaluated using three key metrics: average run length (ARL), standard deviation of run length (SDRL), and median run length (MRL). Simulation results indicate that the proposed resubmission-based adaptive thresholding LASSO control chart outperforms its classical counterpart in detecting disturbances in the covariance matrix. This improvement is particularly evident when simultaneous changes occur in both diagonal and off-diagonal elements. Finally, the practical application of the proposed control chart is illustrated through a case study from the automotive component industry in Iran.

Keywords

High-dimensional data

Resubmission method

Adaptive thresholding LASSO

Covariance matrix

Phase II