Srovnávací analýza metod statistické regulace procesu pro autokorelovaná data.

Abstract

This diploma thesis deals with a comparison of methods for SPC in the case of autocorrelated data. The thesis is based on the comparison of two general approaches to dealing with autocorrelated data: a model - based and a model - free approach. The main part of the thesis is dedicated to the model – free approach and the Batch means method specifically, that is the most developed method of the model – free approach. That part of the thesis is designed as a searching for an alterantive to the model – based methods. The conclusion of the thesis is dedicated to the methods comparison itself. The criteria for the comparison are based on real manufacturing environment implementation, so there were chosen those three criteria: implementation ease, performance in data autocorrelation elimination and the amount of information about process dynamics left.