Control charts can be by variables or by attributes. Most control charts proposed for monitoring univariate process mean use variable inspection. Recent studies propose the use of the ATTRIVAR strategy, which combines the inspection of data through attributes and variables. In the first stage, a sample size n is taken from the process and a go/no go gauge is used to decide whether the second stage, in which the same sample items or items from a new sample will be measured, is required. In stage 1, the np attribute control chart is used and in stage 2, the X variable control chart. The ATTRIVAR strategy is more economical than the inspection only by variable, as most of the time, the second stage is not required. In this project, the performance of new ATTRIVAR control charts for monitoring univariate process mean will be investigated. In the proposed control charts, n items will be evaluated by attributes. In the second stage, only n1 items failed in the first stage will be measured, that is, n1dn. The proposed control charts will be compared with the control charts already proposed in the literature in terms of the speed with which they signal a process disturbance. The properties of the control charts will be obtained by simulation using the R language. The results of the simulation will be used to validate the theoretical results to be obtained by the research group of which the advisor is part.
News published in Agência FAPESP Newsletter about the scholarship: