Economic design approaches of control charts are commonly based on the assumptionthat various cost parameters values and the occurrence risk of assignable causes haveto be a priori known with precision. However, in real operative contexts, such parameterscan be really difficult to accurately estimate, especially considering costs arisingfrom out-of-control conditions of the process. As consequence, pure economic designapproaches can involve chart schemes with low statistical performance. To overcome suchlimitation, it is herein proposed a multi-objective economic-statistical design approachfor an adaptive X-bar chart. In particular, such approach aims at the minimization ofboth the total quality related costs and the out-of-control average run length, in sucha way assuring an optimal trade-off between economic and statistical performance ofthe related control procedure. Moreover, for a robust design approach, the mean shift isconsidered as a random variable. A mixed integer nonlinear constrained mathematicalmodel is formulated to solve the treated problem, whereas the Pareto optimal frontier isdescribed by the ε-constraint method. In order to show the employment of the proposedapproach, an illustrative example is developed and the related considerations are given.Finally, some sensitivity analysis is also performed to investigate the effects of operativeand costs parameters on chart parameters.
|Numero di pagine||23|
|Rivista||INTERNATIONAL JOURNAL OF RELIABILITY, QUALITY, AND SAFETY ENGINEERING|
|Stato di pubblicazione||Published - 2014|
All Science Journal Classification (ASJC) codes