International Journal of Quality Engineering and Technology (6 papers in press)
Distribution-free synthetic and runs-rules control charts combined with a Mann-Whitney chart
by Jean-Claude Malela-Majika, Eeva Maria Rapoo
Abstract: A control chart is one of the most important tools used in statistical process control and monitoring (SPCM) to detect changes in quality processes. This paper investigates the performance of the improved modified distribution-free synthetic and runs-rules charts combined with a Shewhart Mann-Whitney (MW) control charts, in terms of the average run length (ARL), standard deviation of the run length (SDRL) and median run length (MRL) through intensive simulation. It is observed that, the new control charts present very attractive run-length properties and outperform the competing charts in many cases. Numerical examples are given as illustration of the design and implementation of the proposed charts.
Keywords: nonparametric statistical process control; MW statistic; conforming run-length chart; runs-rules; synthetic MW chart; Monte Carlo simulation.
Preventive Maintenance Modeling in Lifetime Warranty
by Farhad Imani, Ki-Hwan Bae
Abstract: Lifetime warranty is a type of long-term assurance that is now
ubiquitous. With a long period of coverage, however, the expected number of
failures and the associated costs are likely to add up. Hence, maintenance
policies over a lifetime warranty can have a substantial impact on warranty
servicing costs. Maintenance policies impose additional costs and would be
worthy only if the reduction amount in total costs due to maintenance is greater
than maintenance costs. The focus of this paper is on modelling preventive
maintenance during a lifetime warranty in order to derive optimal maintenance
policy and optimal level of repair based on the structures of a cost function and
a failure rate function. Our investigation demonstrates that the optimal strategy
for preventive maintenance can be achieved by considering minor repairs
during early ages and marginal repairs near the end of a product life. Numerical
examples are provided to evaluate our developed models and support the
Keywords: Lifetime warranty; increasing failure rate; preventive maintenance policy; failure rate reduction method.
The development of target-based posterior process capability indices and confidence intervals
by Anintaya Khamkanya, Byung Rae Cho, Paul Goethals
Abstract: Quality engineering tools and techniques are often sought as platforms for improving system design, enhancing performance, and optimizing process conditions. Perhaps one of the most popular tools is the process capability index (PCI), which relates the allowable spread of a process defined by engineering specifications to the natural spread of a process. The PCI enables an engineer to assess the performance of a process and thus realize where improvements in product quality may be needed. The vast majority of PCI research involves measuring process performance during the manufacturing stage, prior to the final inspection of products and shipping to the customer. After implementing inspections, however, non-conforming products are typically scrapped when they fail to fall within their specification limits; hence, the actual resulting process distribution shipped to the customer after inspection is truncated. Moreover, the traditional PCI does not account for the loss in quality when product characteristics fail to achieve their process target value. This research, in contrast, proposes indices that consider the underlying result of observations after inspection or when non-conforming products are scrapped, referred to as posterior PCIs. Utilizing a truncated distribution as the basis for measurement along with a target-based quality loss function for capability analyses, several posterior indices are developed corresponding to their traditional non-truncated counterparts. A simulation technique is implemented to compare the proposed posterior PCIs with traditional measures across multiple performance scenarios; finally, the confidence interval approximations for the posterior PCIs are derived. Our results suggest using the proposed posterior indices for capability analyses when industrial processes require that non-conforming products be scrapped prior to shipping to the customer.
Keywords: process capability index; truncated normal distribution; process target; quality loss function.
OPTIMIZATION OF FOOD PROCESSING WITH MULTIPLE QUALITY CHARACTERISTICS USING DESIRABILITY FUNCTION.
by Jesús Gabriel Rangel-Peraza, Edith Padilla-Gasca, Yaneth A. Bustos-Terrones, Jaime Rochin-Medina, Abraham Rodriguez-Mata, Antonio J. Sanhouse-García
Abstract: Desirability function is a statistical methodology that can be applied to find optimal solutions for response variables in multiple objective optimization. Desirability function is widely used for finding a global optimum response for several manufacturing processes, including food processes. In this investigation a factorial design 24 with 4 center points was used to find out the best formulation conditions for the cucumber chutney production process. The factors taken into account were: osmotic dehydration time, thermal treatment time, treatment temperature, and formulation. The response variables measured were: water activity, efficiency, total soluble solids, viscosity and pH. The results showed that it was possible to optimize all response variables in a simultaneous way through the desirability function methodology. An optimization strategy that reduced the food development problem with multiple quality features to a simple mathematical model is the main contribution of this study.
Keywords: Optimization; food processing; quality characteristics; desirability function; design of experiments; response surface methodology; factorial design; desirability function; cucumber Chutney; quality parameters; water activity; efficiency; total soluble solids; viscosity; pH.
Statistical Analysis of the Researches Carried Out on Lean and Six Sigma Applications in Health Care Industry
by Gaurav Suman, D.R. Prajapati
Abstract: The Lean and Six Sigma are two complementary methodologies in the sense that Lean focuses on reducing waste and increasing speed; whereas Six Sigma focuses on reducing variations and increasing consistency. The purpose of this paper is to provide an overview of Lean and Six Sigma applications in the healthcare sector. The work done by many researchers in healthcare industry is discussed. Literature survey shows that most of the studies (42%) are focused on reducing processing time. It is also found that number of studies focused on reducing processing time never goes out of phase. The Pareto chart analysis was performed for number of studies in various countries and different departments. It is found that more than 50% of studies were carried out in United States of America (USA) only and 22% of the studies were performed in emergency department in various countries. The matrix plots are shown for number of studies in different countries and different departments throughout the time line starting from the year 2000 to till date. It is also found that Lean and Six Sigma methodologies were uniformly applied in Emergency and Surgery departments, whereas in case of countries; only USA shows continuous applications of Lean and six sigma techniques.
Keywords: Lean & Six Sigma; Quality in healthcare; Processing time; Productivity; Length of stay.
Hotellings T2 control chart with variable sampling interval and variable dimension
by Reza Shokrizadeh, Mohammad Dolatabadi, Yaqub Yaqubinejad
Abstract: T2 control chart with variable dimension is useful when there is a set of p1 variables that are easy to monitor or whose measurement is cheap, against a set of p2 variables, p = p1 + p2, that are difficult and/or expensive to monitor. However, the information that these p2 variables provide is important to detect quickly the process quality shift. Therefore, there are cases when controlling the whole set of p variables may be difficult or expensive, but controlling sometimes p1 variables, and only when the process seems to have a problem controlling the full set of p variables, may be a cheaper option on average and very efficient. However, its not effective if the shift size is small. For obtaining good performance in detecting such shifts, we propose the application of the variable sampling interval technique to the VDT2 control chart, resulting in the VSIVDT2 control chart.
Keywords: Markov chain; Variable sampling interval; Variable dimension; Average time to signal; Control chart.