drasR

Drift-Aware Risk-Guaranteed Acceptance Sampling

drasR implements Drift-Aware Risk-Guaranteed Acceptance Sampling for adaptive lot disposition under process drift.

The package extends classical acceptance sampling by adding:

• producer-risk and consumer-risk assessment,
• EWMA-based process drift monitoring,
• adaptive inspection tightening,
• Monte Carlo evaluation,
• policy tuning,
• low-cost, balanced, and high-safety policy selection,
• safety-cost frontier visualisation.

Installation

# Development version
# install.packages("devtools")
# devtools::install_github("YOUR-GITHUB-USERNAME/drasR")

Basic example

library(drasR)

plan <- find_plan(
  aql = 0.01,
  ltpd = 0.05,
  alpha = 0.05,
  beta = 0.10
)

plan

Compare fixed acceptance sampling with DRAS

set.seed(123)

result <- run_dras_experiment(
  n_reps = 100,
  scenario = "sudden_deterioration",
  seed = 123
)

result

Tune DRAS policies

tuned <- tune_dras_policy(
  n_reps = 100,
  scenario = "sudden_deterioration",
  seed = 123
)

selected <- select_dras_policy(tuned)

selected

Plot safety-cost frontier

plot_dras_frontier(tuned, selected)

Interpretation

drasR helps users study how adaptive inspection can reduce false acceptance of bad lots under process drift while making the inspection-cost trade-off explicit.

Status

This package is under active development. The current version focuses on simulation-based evaluation and methodological development.