Lifelong Learning for Monitoring and Adaptation of Data-Based Dynamical Models: A Statistical Process Control Approach

The article addresses the stability index of stochastic processes (time series) based on the statistical process control approach. The proposed stability index reveals the relation between assignable and common variabilities: the higher the proportion of the assignable variability, the lower the stability. It is calculated as a ratio of the average moving range to the maximum range between the time series mean and the means of the samples with the fixed end. To address sensitivity of the stability index to process changes such as mean and variance shifts, trends, seasonality, autocorrelation, and sudden changes (outliers), hundred of thousands of simulated time series are analyzed. Considering the impact of sample size, amplitude of process change, moment of process change, and type of distribution, the full factorial design of experiment has been applied. The proposed calculation of the stability index and suggested target levels serve as a guide for interpreting stability of the stochastic processes by index. A combination of stability and capability indexes emphasizes dynamic aspects of the process ability regarding specifications and are presented as a dynamic capability index.

Although the traditional control algorithm produces many successful results on system integration applications, there still exist insufficiencies need to be improved. For traditional control approach, it is well known that the search equation is superior at local range but constrained at global range for system integration, which will transform the optimal performance of the algorithm. Therefore, the search ability is the most important goal in algorithm modification on system integration. To solve the faults in traditional control algorithm and achieve the goal of improvement, a new statistical process control approach is proposed to global optimization of system integration. In order to balance the search ability, search algorithms are made better by search strategies that search probability and random variable is related with a certain degree of correlation. In addition, to deepen the convergence degree, the local search operator is also utilized. The simulation results tested show that the proposed process control approach can outperform control approach in most of the experiments.

The classical funnel experiment was used by Deming to promote the idea of statistical process control (SPC). The popular example illustrates that the implementation of simple feedback rules to stationary processes violates the independence assumption and prevents the implementation of conventional SPC. However, Deming did not indicate how to implement SPC in the presence of such feedback rules. This pedagogical gap is addressed here by introducing a simple feedback rule to the funnel example that results in a nonlinear process to which the traditional SPC methods cannot be applied. The proposed method of Markov‐based SPC, which is a simplified version of the context‐based SPC method, is shown to monitor the modified process well. Copyright © 2007 John Wiley & Sons, Ltd.

PurposeThe purpose of this paper is to develop a new easy-to-implement distribution-free integrated multivariate statistical process control (MSPC) approach with an ability to recognize out-of-control points, identify the key influential variable for the out-of-control state, and determine necessary changes to achieve the state of statistical control.Design/methodology/approachThe proposed approach integrates the control chart technique, the Mahalanobis-Taguchi System concept, the Andrews function plot, and nonlinear optimization for multivariate process control. Mahalanobis distance, Taguchi’s orthogonal array, and the main effect plot concept are used to identify the key influential variable responsible for the out-of-control situation. The Andrews function plot and nonlinear optimization help to identify direction and necessary correction to regain the state of statistical control. Finally, two different real life case studies illustrate the suitability of the approach.FindingsThe case studies illustrate the potential of the proposed integrated multivariate process control approach for easy implementation in varied manufacturing and process industries. In addition, the case studies also reveal that the multivariate out-of-control state is primarily contributed by a single influential variable.Research limitations/implicationsThe approach is limited to the situation in which a single influential variable contributes to out-of-control situation. The number and type of cases used are also limited and thus generalization may not be debated. Further research is necessary with varied case situations to refine the approach and prove its extensive applicability.Practical implicationsThe proposed approach does not require multivariate normality assumption and thus provides greater flexibility for the industry practitioners. The approach is also easy to implement and requires minimal programming effort. A simple application Microsoft Excel is suitable for online implementation of this approach.Originality/valueThe key steps of the MSPC approach are identifying the out-of-control point, diagnosing the out-of-control point, identifying the “influential” variable responsible for the out-of-control state, and determining the necessary direction and the amount of adjustment required to achieve the state of control. Most of the approaches reported in open literature are focused only until identifying influencing variable, with many restrictive assumptions. This paper addresses all key steps in a single integrated distribution-free approach, which is easy to implement in real time.

ObjectiveSurgical site infections (SSIs) are common costly hospital-acquired conditions. While statistical process control (SPC) use in healthcare has increased, limited rigorous empirical research compares and optimises these methods for SSI...

This paper introduces an uncertainty model being developed for the National Transonic Facility (NTF). The model uses a Monte Carlo technique to propagate standard uncertainties of measured values through the NTF data reduction equations to calculate the combined uncertainties of the key aerodynamic force and moment coefficients and freestream properties. The uncertainty propagation approach to assessing data variability is compared with ongoing data quality assessment activities at the NTF, notably check standard testing using statistical process control (SPC) techniques. It is shown that the two approaches are complementary and both are necessary tools for data quality assessment and improvement activities. The SPC approach is the final arbiter of variability in a facility. Its result encompasses variation due to people, processes, test equipment, and test article. The uncertainty propagation approach is limited mainly to the data reduction process. However, it is useful because it helps to assess the causes of variability seen in the data and consequently provides a basis for improvement. For example, it is shown that Mach number random uncertainty is dominated by static pressure variation over most of the dynamic pressure range tested. However, the random uncertainty in the drag coefficient is generally dominated by axial and normal force uncertainty with much less contribution from freestream conditions.

Visualizing and understanding the underlying processes in short production run environments is not possible with standard Shewhart control charts. Administration of multiple charts, frequent setups and changeovers, and the truncated nature of data collection limit the effectiveness of standard statistical process control (SPC) approaches. This article introduces a process visualization approach by adapting short-run SPC and recommends new rules for interpretation.

In spite of the recent advancements in wafer fabrication techniques, devices are still being individually checked and modified. This is due to the fact that the general manufacturing processes involved in wafer production have their inherent inconsistencies. As a consequence, individual devices show differences in characteristics that would render a big group of them operating out of the range of the pre-set spec limits. Therefore they would require types of modifications specific to the individual device. Knowing that the resources spent on checking and eliminating out-of-spec devices before they reach the customer are very significant, the manufacturing operation becomes hardly profitable. These wafers normally carry devices in the range of a thousand or so, making a statistical approach very attractive. In this paper, an actual industrial problem in wafer fabrication to the desired specifications is presented. The problem shows in passive filters built using Surface Acoustic Wave (SAW) theory. A solution considering a statistical process control approach to the population of devices on the wafer is proposed. The results of applying this solution are realized in significant product yield increase, huge cost cutting, and automation promotion and application.

Technological advancements have profoundly influenced various sectors, including tourism, by simplifying international travel and increasing the demand for passports. In this context, public feedback on the Surabaya Immigration Office's services, gathered from Google Maps reviews, presents an invaluable dataset for analysis. This study introduces a novel approach to quality management systems in immigration services by applying the Convolutional Long Short-Term Memory (Co-LSTM) for sentiment analysis of public reviews and p-attribute control charts for statistical process control. Focusing on the Surabaya Immigration Office, we analyze public feedback from Google Maps to assess service quality. Our methodology preprocesses and classifies opinion data into sentiment classes, employing Co-LSTM for its superior accuracy over traditional models. The sentiment analysis reveals a predominance of positive over negative reviews, with classification accuracy demonstrating Area Under Curve (AUC) values of 98.61% for training and 85.66% for testing data. Furthermore, the p-attribute control charts are utilized to monitor service defects, identifying areas of uncontrolled variability and suggesting the necessity for service improvement interventions. The study uncovers that the primary public grievance relates to the perceived lack of friendliness and politeness from office staff. By integrating sentiment analysis with statistical process control, this research offers a comprehensive approach for immigration service providers to enhance service quality, respond proactively to public sentiment, and ensure customer satisfaction.

Received: 29 October 2012 Abstract Accepted: 28 November 2012 Traditional statistical process control approaches are less effective in dealing with multivariate and autocorrelated processes. With the continual increase in process complexity, this inefficiency is becoming more apparent. A special type of multivariate and autocorrelated process is a process occurring within a heterogeneous production environment (a variety of types of machines, pots, etc. used for the same task). This makes the quality control of such processes more difficult. The approach presented in the paper utilizes time series fitting, cluster analysis and association mining in relation to a single data mining model for the analysis of complex multivariate autocorrelated processes. The aim is to divide the production cells (machines, pots, etc.) into groups exhibiting similar behaviors. This can then be used for more effective quality control of the entire process and afterwards to analyze the reasons for this behavior. This paper includes someof the results obtained from applying the model to an actual multivariate high autocorrelated process, the production of primary aluminum using the Hall-Heroult electrolysis process. The Hall-Heroult electrolysis process is a continual process that is ongoing in several pots simultaneously. The average plant operates 300 pots. Therefore, the quality control of such a complex process faces many issues concerning monitoring and problem diagnosis. The paper describes a method for dividing the pots into control groups exhibiting similar behaviors, which can then be used in the planning phase of the quality control analysis and to make improvements within these groups and thereby within the whole process.

Based on the bridge health monitoring data, the traffic-load effect model for arch bridge suspender was set up using direct statistical method. Firstly, the Bridge Health Monitoring System (BSHM) set up on the Ebian Dadu River Bridge in Sichuan province to obtain daily traffic-load effect of suspenders was introduced. Secondly, the section distribution of suspender traffic-load effect was researched by statistical approach using monitoring data. Finally, the probability distribution of maximum traffic-load effect during design period was given. The statistical result shows that the section distribution of suspender traffic-load effect follows lognormal distribution and the probability distribution of maximum traffic-load effect during design period follows generalized extreme-value distribution. The traffic-load effect model of suspender was different with the design code given which may underestimates the real load effect level. It's the first time in China to determine the traffic-load effect probability model for arch bridge suspenders based on actual acquired monitoring data using statistical and probabilistic approach.

Prediction Control Charts: A New and Flexible Artificial Intelligence-Based Statistical Process Control Approach

In the era of digital transformation, the integration of artificial intelligence within enterprise resource planning (ERP) systems is redefining financial management. This research explores the unprecedented application of generative AI modelssuch as transformer-based architectures-for real-time anomaly detection within SAP FICO modules. Unlike traditional rulebased or statistical approaches, generative AI can learn complex, dynamic financial patterns, enabling the proactive identification of irregularities in general ledger entries, accounts receivable, and accounts payable. This study proposes a novel framework that embeds generative AI directly into SAP FICO workflows, facilitating continuous monitoring, adaptive learning, and automated escalation of financial anomalies. The research also addresses data privacy, explainability, and integration challenges, offering a roadmap for organizations seeking to enhance their financial governance and compliance posture. Through simulation and case-based analysis, the paper demonstrates how this approach can reduce financial fraud, improve audit readiness, and drive operational efficiency-heralding a new era of intelligent financial oversight in SAP environments.

Studies show that an adaptive learning environment is needed by learners. Adaptive e-learning systems have been developed with personalized instructions based on student's knowledge level. Since each student has a unique cognitive and metacognitive level, we currently develop an adaptive Hypermedia Learning Environment (HLE) for Algorithm and Data Structure Course based on a metacognitive awareness level. HLE provides adaptive scaffolding as instructional interventions to facilitate students' aptitude in programming. Scaffolding can be determined by knowing student's metacognitive awareness level. In this research, we propose the use of Tahani Model of Fuzzy Database in web-based HLE to support the adaptive decision-making. Tahani is one of some feasible models in a fuzzy database that can categorize students' metacognitive awareness level. The produced fuzzy rules have been validated by metacognitive experts of educational psychology. A hundred students have been categorized into three groups (low, medium, high) and provided with appropriate scaffolding. Evaluation results verified the suitability of the proposed method and the student's scaffolding.

Passivity-Based Iterative Learning Control for Spacecraft Attitude Tracking on SO(3)