A linear-time method for computing the ANOS of Bernoulli and Poisson CUSUM charts.

The average number of observations to signal (ANOS) is an important measure of the effectiveness of Bernoulli or Poisson cumulative sum (CUSUM) control charts. Being able to quickly and accurately calculate an ANOS vector facilitates effective control chart design. We present a linear-time method for computing the ANOS of Bernoulli CUSUM charts and generalize to Poisson CUSUM charts. This method overcomes computation challenges associated with previously existing methods.