1121 - Alexey Antonitsin

Statistical Methods in Reliability Testing


In this project we examine different sampling plans as part of reliability control procedures in manufacturing. The systematic sampling conventionally used by many companies can be replaced by a skip-lot sampling plan which provides the same level of protection against defective products while requiring fewer resources. In this project we investigate skip-lot sampling procedure and its statistical properties. We will also compare the medium rank regression method of fitting Weibull distribution parameters with the maximum likelihood method. Finally, we conduct a simulation study to empirically investigate the performance of the proposed method and its properties. In contrast to the existing literature on this topic, this project considers a wider range of reliability sampling procedures and makes extensive use of simulation methods.