Binomial and trinomial versions of the ATTRIVAR control chart

The ATTRIVAR control chart uses attribute and variable sample data to control the process mean of a quality characteristic X. The trinomial version of the ATTRIVAR chart, presented in this paper, uses, firstly, two discriminating limits to classify the sample items into three excluding categories. Items can be classified in categories A, B and AB in which the items have the X dimension, respectively, lower than the lower discriminating limit, greater than the upper discriminating limit and between the lower discriminating limit and the upper discriminating limit. The number of sample items in each category lead to one of three decisions: the process is stated as in-control, the process is stated as out-of-control or the X dimension is measured for all the sample items. If the decision made is this last one, the sample mean of the X values is used to decide the state of the process. The binomial version classify the items into two excluding categories, defective (A and B of the trinomial) and not defective (AB of the trinomial). The value of Average Run Length (ARL) were used to evaluate the efficiency of this control chart. The trinomial version of the ATTRIVAR chart is more efficient than the binomial version in detecting a shift in the process mean.