IECM 12.0 beta User Manual > Using the IECM > Analysis Tools > Uncertainty > Configure > Sampling Method >

Random Sample (Monte Carlo)

The Random Sample sampling method is also known as Monte Carlo. Monte Carlo is the simplest and best-known sampling method. It draws values at random from the uncertainty distribution of each input variable in the decision tree. For a particular sampling run, each input variable is randomly sampled once. The random samples from each input result in one final output value. This process is repeated m times and results in a final solution set. This set can then be evaluated with standard statistical techniques to determine the mean, precision, and confidence.

This method has the advantage of providing an easy method of determining the precision for a specific number of samples using standard statistical techniques. However, it suffers from requiring a large number of samples for a given precision. It also has the drawback of substantial noise in the resulting distribution. For these reasons, Latin Hypercube sampling is preferred as the model default.