Many processes are so complex that determining which factors are important is very difficult. One approach is to simulate the process. Using a computer simulation, the process may be run many times. Via descriptive statistics, important factors may be found. Their significance may be determined from the simulations. The simulations may be run varying the basic parameters, e.g. have employees work overtime or using temporary workers.
By use of "bootstrapping" techniques, confidence or prediction intervals may be constructed with no assumptions of the underlying probability distributions.