Stopping Monte Carlo

The Monte Carlo errors are important to define stopping rules. The traditional engineering idea is that the computing errors should not be greater then the data gathering errors. Considering the call centers the data errors usually means the call rate prediction errors. Thus the stopping rule is:
- stop at some observation $k \ge K$ defined by this inequality:

$$\sigma(K) < \sigma$$ (95)

where $\sigma$ is a standard deviation of call rate predictions.