Quality improvement prioritization in production systems
Betterton, Carl E
MetadataShow full item record
Much effort and money are wasted in misguided "improvement" efforts because those involved lack effective methods to choose among improvement alternatives. This dissertation documents an investigation of how to prioritize quality improvements in production systems, where improvement is understood to be a reduction of the scrap rate at one or more workstations within a system. In a manufacturing organization, yield loss (scrap) directly and negatively affects individual workstations as well as overall production system performance. Once a production system is in operation, an important factor in dealing with yield loss and its performance impact is planned improvement efforts. To make effective decisions about application of this factor, management must understand how a system will respond to an improvement approach. This implies an understanding of important system differences as well as an appreciation for the correct approach. Systems differences of interest in this study were product structure and constraint location. Several different production configurations, i.e., product structure – constraint location combinations, were investigated. To the author’s knowledge, this is the first research to make systematic comparisons of quality improvement across product structures. In addition, there is very little published research showing how constraint (bottleneck) location influences the effects of improvement. Four approaches or decision rules to guide improvement were defined. Three were derived from the quality literature and were generally exercised within the context of traditional accounting performance measures. The fourth used a holistic or systems approach, and relied on performance measures proposed in the Theory of Constraints (TOC) literature. The approaches are, for a given system configuration: 1. Choose the workstations to improve based on the number of defective items (scrap) produced. 2. Choose the workstations to improve based on the dollar value of defective items (scrap) produced. 3. Improve all workstations across the board; all workstations have equal priority for improvement. 4. Choose the workstations to improve based on TOC principles. Discrete-event simulation experiments were performed based on a factorial design that included the four prioritization rules and three different constraint locations for each of the two most common product structures. Each combination of product structure and constraint location represents a unique system configuration. Application of a given prioritization rule on a specific system configuration represents one treatment. On each of the six unique line configurations, four independent treatments were applied, distinguished by the prioritization rule that guides them. Each treatment (improvement) was applied repeatedly through four cycles of three months each, covering one full year. At the end of the year, the values of the two system response variables (good product and scrap product) for each treatment were observed and compared to the values of these variables in the unimproved system. Results were analyzed using analysis of variance (ANOVA) with Tukey multiple comparisons. Results of this investigation show that constraint location has a major impact on profit and that approach 4 (TOC) results in maximum system improvement and profit. Regarding preferred constraint location, this study, which included yield loss and concluded the best location was near the end of the line, contradicts all known previous research, which assumed perfect quality and concluded the best location was near the front of the line.