LGCU (version 0.1.5)

plot_GICCLup_Chart: Upward CUSUM Control Chart with Cautious Learning and Guaranteed Performance

Description

This function generates an upward CUSUM control chart for a Gamma distribution, incorporating a cautious parameter update mechanism based on guaranteed performance.

It enables dynamic process monitoring, ensuring progressive adaptation to distribution changes. This approach follows the methodology presented in the work of Madrid-Alvarez, García-Díaz, and Tercero-Gómez (2024), where a cautious learning scheme for parameter updates in CUSUM control charts applied to Gamma distributions is proposed.

The implementation captures distribution changes and adjusts the control limits to improve process variation detection.

Features:

• If the user does not provide Phase I and Phase II data, the function automatically generates them.

• If known_alpha = TRUE, alpha is fixed and not estimated.

• If known_alpha = FALSE, alpha is estimated from Phase I data.

• Includes dynamic control limits and a summary of parameters.

• Integrates a cautious learning scheme using the parameters k_l, delay, and tau.

Recommendations

The parameters k_l, delay, and tau are part of the cautious learning mechanism of the CUSUM chart. These values enable the dynamic updating of beta0_est and H_plus, ensuring that the control chart progressively adapts to process changes, thus improving sensitivity in detecting deviations.

For proper implementation, it is recommended to reference the values proposed in:

Madrid-Alvarez, H. M., García-Díaz, J. C., & Tercero-Gómez, V. G. (2024). A CUSUM control chart for the Gamma distribution with cautious parameter learning. Quality Engineering, 1-23.

While these parameters have been tested and validated in the referenced article, users can adjust them based on the specific characteristics of their process, considering factors such as system variability and desired update frequency.

Additionally, if detailed guidance on selecting values for H_delta and H_plus is required, it is recommended to review the referenced article, which presents calibration and adjustment strategies for these limits to ensure optimal control chart performance.

Usage

plot_GICCLup_Chart(
  alpha,
  beta,
  beta_ratio,
  H_delta,
  H_plus,
  known_alpha,
  k_l,
  delay,
  tau,
  n_I,
  n_II,
  faseI = NULL,
  faseII = NULL
)

Value

A plot showing the evolution of the upward CUSUM statistic with cautious learning, including:

• The dynamically adjusted accumulated values of the CUSUM statistic.

• Progressively updated control limits with guaranteed performance.

• A summary of the parameters used in the control chart.

Arguments

alpha

Shape parameter of the Gamma distribution (if known_alpha = TRUE).

beta

Scale parameter of the Gamma distribution.

beta_ratio

Ratio between beta and its estimation.

H_delta

Increment of the upper control limit.

H_plus

Initial upper control limit of the CUSUM chart.

known_alpha

Indicates whether alpha is known (TRUE) or should be estimated (FALSE).

k_l

Secondary control threshold used in the learning logic.

delay

Number of observations before updating beta0_est and H_plus_c.

tau

Time point at which the beta parameter changes.

n_I

Sample size in Phase I (if faseI is not provided).

n_II

Sample size in Phase II (if faseII is not provided).

faseI

Data sample from Phase I (numeric vector). If NULL, it is generated internally.

faseII

Data sample from Phase II (numeric vector). If NULL, it is generated internally.

Examples

# Option 1: Providing Phase I and Phase II data
phaseI_data <- rgamma(n = 200, shape = 1, scale = 1)
phaseII_data <- rgamma(n = 710, shape = 1, scale = 1)
plot_GICCLup_Chart(alpha = 1, beta = 1, beta_ratio = 2, H_delta = 4.2433,
                   H_plus = 8.9345, known_alpha = FALSE, k_l = 2, delay = 25,
                   tau = 1,n_I = 200, n_II = 700, faseI = phaseI_data,
                   faseII = phaseII_data)

# Option 2: Without providing data, the function automatically generates them
plot_GICCLup_Chart(alpha = 1, beta = 1, beta_ratio = 2, H_delta = 2.9819,
                   H_plus = 6.5081, known_alpha = TRUE, k_l = 2, delay = 25,
                   tau = 1, n_I = 200, n_II = 710, faseI = NULL,
                   faseII = NULL)