A high‐dimensional control chart for profile monitoring

Profile monitoring is an important and rapidly emerging area of statistical process control. In many industries, the quality of processes or products can be characterized by a profile that describes a relationship or a function between a response variable and one or more independent variables. A change in the profile relationship can indicate a change in the quality characteristic of the process or product and, therefore, needs to be monitored for control purposes. We propose a high‐dimensional (HD) control chart approach for profile monitoring that is based on the adaptive Neyman test statistic for the coefficients of discrete Fourier transform of profiles. We investigate both linear and nonlinear profiles, and we study the robustness of the HD control chart for monitoring profiles with stationary noise. We apply our control chart to monitor the process of nonlinear woodboard vertical density profile data of Walker and Wright (J. Qual. Technol. 2002; 34:118–129) and compare the results with those presented in Williams et al. (Qual. Reliab. Eng. Int. 2007; to appear). Copyright © 2010 John Wiley & Sons, Ltd.     

A forecast‐based monitoring methodology for process transitions

Nonconforming parts are often produced when a process moves from one level to another due to transition events. Control charting, when applied to a stable state process, is an effective monitoring tool to continuously check for process shifts or upsets. However, the presence of transition events can impede the normal performance of traditional control chart with increased false alarms. The presence of autocorrelation also requires modification to the control charting procedure. We present a methodology for characterizing the process transition which involves a tracking signal statistic, based on the forecast‐based exponentially weighted moving average (EWMA). This test will supplement the forecast‐based EWMA control charting as a means of detecting when the transition event is complete. Such a procedure facilitates smooth application of the appropriate control chart by knowing when the transition is over. The transition characterization methodology also carries benefits in cost and material savings. We use a color transition process in plastic extrusion to illustrate a transition event and demonstrate our proposed methodology. Simulation is employed to evaluate the performance of the methodology. Copyright © 2001 John Wiley & Sons, Ltd.     

Cross Training in Serial Production With Process Characteristics and Operational Factors

Worker multifunctionality has been recognized as a tool for enhancing system flexibility and performance. This paper investigates the impacts of worker multifunctionality in a heterogeneous serial production system in the presence of individual differences in experiential learning and forgetting. The tradeoffs of multifunctionality and cross training are examined in conjunction with several process characteristics and operational decisions including staffing level, system bottleneck position, task similarity degree, and worker rotation interval. The results show the relationships between the preferred level of worker multifunctionality and both process characteristics and operational decisions. In several cases, minimum multifunctionality and short worker rotation intervals provide improved system performance. Additionally, process characteristics, such as the position of the system bottleneck and the degree of task similarity, have impacts on individual learning--forgetting behaviors and, subsequently, on decisions relating to multifunctionality levels and system performance.     

Simulation using the state-space representation of noisy dynamic systems to determine effective integrated process control designs

There has been considerable interest in investigating control policies that can be applied to dynamic systems to take advantage of engineering process control (EPC) and statistical process control (SPC). We design two simulation models using SIMULINK to represent control of noisy dynamic systems given by a state-space representation. We use the models on systems with dynamic behavior that have been degraded by an additive ARMA noise process. Our findings show that the combination of EPC and SPC is a more effective policy on a complex dynamic system than on a simple dynamic system.     

Rule mining for scheduling cross training with a heterogeneous workforce

This paper presents an association rule mining based framework for workforce scheduling to assist managers with robust real-time assignment decisions. We assume heterogeneous individual learning and forgetting behaviours, in which worker productivity changes dynamically. We explore a parallel production system that meets a specified production requirement over a fixed time horizon with the minimum workforce resources based on the number of worker-periods assigned. Three managerial policies are considered including: setting a maximum allowable individual cross-training level, balancing workload among workers and an unconstrained policy. We propose the use of several schedule attributes to quantify key aspects of optimised schedules that may, in turn, aid in determining robust assignment rules and the development of better cross-training policies. Current results indicate that the proposed approach is effective at identifying important rules, many of which add to our knowledge of useful workforce scheduling strategies.     

The effects of workforce heterogeneity on productivity in an experiential learning environment

This paper investigates the effects of heterogeneity in various workforce learning/forgetting characteristics on productivity. An important aspect of this study is that worker heterogeneity is specifically modeled and examined by including the variance of key parameters as controlled experimental factors in addition to these parameters’ means. Initially, we analytically investigate the impact of heterogeneity in each of the four parameters in the learning/forgetting model used in this study. The analytical results are subsequently verified and further investigated using computer simulation. Specifically, seven macro-level factors are investigated including the rate of turnover, the mean and variance of the latent initial expertise existing in the workforce for the tasks to be performed, the mean and variance of worker steady-state productivity levels, the variance of the rate of worker learning, and the variance of working forgetting. The simulation models were developed and informed by empirical data from industry. The results indicate significant effects for mean initial expertise, mean steady-state productivity, the variance of initial expertise, the variance of the rate of worker learning, and the variance of worker forgetting. Results have managerial implications in the comparison and selection of various workforce populations for production systems.     

Authors' reply to "A note on 'An empirical comparison of forgetting models'"

The modeling of human learning and forgetting has justifiably received increased attention in recent years. The ability to bridge the gap between the findings of experimentalists and empiricists in this area should allow for the research community to develop ever-improving models. In their research note Jaber and Sikstrom (see ibid., vol.51, no.2, p.233-4, 2004) raise an important issue, namely the question of how well the learn-forget curve model (LFCM) can be expected to perform in practical situations. Past studies have shown a range of expectations regarding this model. From a definitive standpoint, while the present discussion may not fully resolve the question, there are significant and interesting differences between these studies, which are described below. The goal of this note is to clarify and quantify these differences in order to set a framework for continued work on the LFCM model and potentially other models in this area.     

Structural models of extraversion,communication, and team performance

We consider the degree to which the Five Factor Model characteristic of individual extraversion is predictive of the overall performance of dyadic teams. An Anti-Air Warfare Coordinator (AAWC) simulation is employed in an experimental environment wherein dyads are required to communicate key information to complete a series of tasks. Communication is monitored with respect to the number of utterances, number of words per utterance, and the durations of each utterance. Several structural equation models are considered to examine the direct and indirect effects of individual extraversion on team performance. Notably, the degree of extraversion did not directly affect team performance. However, there are several models that indicate significant relationships between extraversion and team performance with communications as mediators. The current results contribute to the literature by adding to our understanding of the mediating variables that exist within the personality–performance team dynamic.     

Transition Monitoring and Adjustment for Dynamic Systems in a Process Improvement Environment

In manufacturing applications, we often encounter process transitions due to a changeover in the production or perhaps an unknown perturbation. The main process improvement goal is to shorten the transition time by monitoring the process in order to quickly identify the start and end of the transition period and by actively adjusting the process during the transition. To address these issues, we propose a transition monitoring and adjustment methodology. A polymer process is used to illustrate this methodology. Using simulation, we characterize the impact of the transition adjustment on the effectiveness of monitoring. We show that the adaptive monitoring procedure is robust to small transition adjustments, thus supporting a complimentary application of process monitoring and process adjustment to improve process transitions. Copyright © 2005 John Wiley & Sons, Ltd.     

Model comparison in Emergency Severity Index level prediction

Emergency Department (ED) triage is a process of determining illness severity and accordingly assigning patient priority. The Emergency Severity Index (ESI) is a 5-level acuity categorization system that aides in triage. This paper compared the capabilities of predicting ESI level using ordinal logistic regression (OLR), artificial neural networks (NNs), and naïve Bayesian networks (NBNs). Data were obtained from Susquehanna Williamsport Hospital for 947 patients over a one month period in 2008. It contained the assigned ESI level, chief complaint, systolic blood pressure, pulse, respiration rate, temperature, oxygen saturation level (SaO₂), age, gender, and pain level. An OLR model was fit using a subset of these covariates. NBNs and NNs were modeled to relax the inherent assumptions of linearity and covariate independence in logistic regression. These three techniques were compared using incremental training dataset sizes between 50% and 100% of given data. All models were >60% accurate using the entire dataset for training. It was found that NBNs and NNs were robust to data size changes and all models had evaluation speeds of less than 0.5 s. At this time the use of NBNs is recommended considering speed, accuracy, data utilization, model flexibility, and interpretability of the model.