Capacity of a process from

from a graph by attributes

1DEFINITION
The process capability is defined as the level of variation that makes it
it meets or not the established requirement. This variability is
related to the natural fluctuations that exist in the process or
inherent variations to it.
In the quality glossary, process capability is defined as 'a
statistical measure of the process to the inherent variability of a
given characteristic. The most widely used formula accepted for the
The process capability is 6 sigma.” (American Society for Quality, 2005).
This means that processing capacity refers to the moment in which
that the process is free from assignable variation and operates with the
unassignable variability or natural variability. This natural variation does not
a factor is assigned that affects control such as poor raw material, a
inexperienced worker, a procedure, a cumbersome measurement, a
misaligned machinery, or an unpredictable environment. In summary
after the Six M's (machinery, methods, labor, material
first, environment and measurements), are completely found
controlled, a certain degree of continues to be present in the process
variation, this is the unassignable variation.
One way to calculate process capability is through 'the indices
of capacity, which will be determined by the ratios between the variation
"birth of the process and the level of specified variation." (Open University of
Catalonia, 2012
The process capability indices are the ratio of the limits of
specification and the inherent variation of the process. In itself, "The indices of
process capability relates to the natural tolerance of the process and to the
engineering specifications. In most control situations
quality, the natural tolerance is set in terms of 6 sigma, while the
engineering specifications are set independently of the
parameters of the distribution." (Kocherlakota, 1992)
2 PROCESS CAPACITY IN ATTRIBUTES
La medición de la capacidad de proceso es una técnica cuantitativa que
it has usually been applied using the reference characteristics of
quantitative quality or variable type. According to Douglas Montgomery, "The
process capability can be expressed as a percentage outside of the
specifications. This percentage that is outside of the specifications is
it can be considered as the defective fraction of the control charts P and NP.
For attribute control charts, it is not customary to apply
the capacity indices, although the following expressions are known
as an informal way of 'calculating' capacity.
2005
For Control Chart P:(1-P) * 100%
For control chart np: n-np/n * 100%
For control chart c: c
For control chart u: u
For the P graphics, the complement of P, which is 1-P, should be used, just like
way for np. For the case of control charts C and U, capacity
of the process is given by the arithmetic averages that use the graphs
of control that are the values of C and U respectively.
"The study of process capability by attributes is reflected in terms of
production of the defective fraction and defects per unit of
production. The main tools used in the studies
Process capability in attributes is the control charts for attributes.
(Gitlow, Levine, & Popovich, 2006)
Following the same scheme as the control charts for attributes 'The
process capability can be estimated in terms of errors or
mistakes instead of the variability of a process parameter.
(Gryna, Chua, & Defeo, 2007) Control charts for attributes, in
especially the control charts U and C, are based on the distribution of
Poisson probability, it has a parameter lambda, λ = defects per
unit or error per unit. So the process capability can be given
by
Defect
Process capability=
Physical unit

A physical unit is understood to be an employee, a production unit, a product,

a service process, among others. Where capacity can be given in
terms of:
 Errors by operator
Defects per unit
Defects by product
Errors in a banking transaction
In summary, it is possible to have processing capacity in attributes.
in terms of sigma level, in defective fraction or the complement which is
performance, errors or defects.
This is why it is important to take into account the reduction of variability.
and also the reduction of the non-compliant fraction in the processes,
sharing the affirmation of Pearn and Wu, which says 'In modern philosophy
Nowadays, improvement in quality is based on reduction of the
deviation with respect to a target value and is as important as the
reduction of the defective units fraction." (Pearn & Wu, Variables
sampling plans with PPM fraction of defectives and process loss
consideration, 2006) That is why it is established that for a study of
process capability should be established in these two ways of measuring the
process capability, initially to reduce variability and on the other hand
also the fraction of defective units. In other words
when working with quantitative data it is ideal that they meet
a target value, while for attributes the achievement is that they do not occur
defective units. This is illustrated in the following graph:

3 TECHNIQUES FOR MEASURING CAPACITY

PROCESSES FOR ATTRIBUTES
To work on measuring process capability with characteristics of
quality type attribute, it should be taken into account that the values that can be
obtain in an attribute are dichotomous (when they can only adopt one single
value without hierarchy among them: man - woman, positive - negative, present -
absent), for the case of quality it is good or bad product, this takes a
value, not two values at once. In this sense, the evaluation of quality in
attributes are based on defects and products or services
defective.

3.1 TECHNIQUES FOR DEFECTIVE

To evaluate the process capability in attributes, based on the
identification of the number of defective items or the defective fraction, is
the following techniques are found:
Process capability indices for taxes
Process capability indices for qualitative data, based on the
Taguchi quality loss function
3.1.1 Process capability indices for attributes (Lluis Cuatrecasas)
Just like the management of process capacity indices in variables,
it must start from the attribute control charts, controlled
Statistically, Lluis Cuatrecasas designs indicators similar to those of CP.
and CPK. (Cuatrecasas, 2005)
The focus on the measurement of process capability in attributes
it is generally done for services, its rating is based on two
parameters, acceptance or non-acceptance of the service. Taking into account the
process capability index equation, the control limits for
defective units and the Cp index; Cuatrecasas derives the equation for
evaluate the process capacity in attributes like this:

Starting from equation (2), it is taken into account that in attributes Ll = 0, and
taking into account the standard deviation of the control charts by
attributes, equation (3) is obtained for the calculation of Cp in attributes.

This capacity index, according to the result it yields, is analyzed and

interpret it this way:
CP < 1 Incapable process
CP = 1 Justly capable process
 CP > 1 Capable process
As this indicator approaches 2, it is a level six process.
sigma.
For the case of the Cpk index, which is the actual capability index of a process,
this takes into account the centering of the process, it is also an adjustment of the index
Cp, it is established that the equation of the Cpk index is:

From the equations of the attribute control charts np or

equation (1), the parameter σ of the control chart

√
NP= np(1−
np
n
) and the

parameter μ=np, from equation (4) of the Cpk index, the

parameters and the process capability index equation is obtained from
CPk for measuring process capacity in attributes:

For most cases Ll = 0, which is an ideal value of the process.

(Six Sigma quality), also remember that LS is the acceptable value of
process defects.
3.1.2 Process capability indices for qualitative data, based on
the Taguchi quality loss function (Hsieh and Tong)
This technique emerged from relating the process capability indices (CPI)
with the theory of Taguchi's quality loss function. (Hsieh &
Tong, 2006) Before starting with the mathematical development to obtain
the process capability indices, it is important to clarify that the limit of
the specification used for qualitative data is unilateral and the goal is
produce zero defects.
The quadratic quality loss function of Taguchi is given by:
L(y) = k(y - T)2(6)
The expected lost quality is given by:
QL(y) = E[L(y)] = E[k(y – T)2]
QL(y) = E[k(y - μ + μ - T)2] (7)
Where (y - μ)2= σ2, therefore it is obtained:
QL(y) = [k((μ – T)2+ σ2) (8)
Where:
Y= corresponds to the data
T denotes the target value or nominal value
K= denotes the constant of quality loss when the process is
finds within the permissible tolerance
μ = denotes the mean of the process

σ denotes the deviation of the process

In accordance with the corollary that: "For qualitative data, the situation
the objective is zero defects (T = 0), meaning that none occur.
"discrepancies." (Hsieh & Tong, 2006) Therefore, having this condition
from equation (8), it follows:

QL(y) = [k(μ2+ σ2] (9)

Considering that the measurement of process capability in variables
is given by the ratio between the customer's specifications and the variation
of the process; according to Taguchi's theory this would be the loss of quality of
client, permissible regarding the actual loss of quality. Therefore, the index of
process capability index (CPI) for attributes can be generalized in the
next expression:

Where

Una de las distribuciones de probabilidad que siguen los defectuosos está

given by the binomial distribution, it has parameters μ= p and σ2=
p(1-p), replacing we obtain:
Finally, it is obtained that the process capability is the relationship of
the defective fraction that the customer is willing to accept regarding
the defective fraction of the process is defined by the following expression:
Pc
CPI= (12)(Hsieh & Tong, 2006)
P