Describing Monte Carlo tolerance analysis output

Question

I am trying to perform some Monte Carlo tolerance analysis, but I am having trouble interpreting what some of the output data means. Can you please indicate where I can find an explanation of the Monte Carlo tolerance analysis output?

Synopsis

Describing Monte Carlo tolerance analysis output

Solution

"The OSLO on-line help system is the recommended place to go for detailed explanations of the various analyses contained in OSLO. The on-line help is easily updated with each release so that it is considered to the most up-to-date and accurate description of the program features. In this particular case, the output of the Monte Carlo tolerance analysis is described in the following online help section:"Contents>>Tolerancing>>Monte Carlo Tolerancing" The data that commonly appears in standard statistical analyses is described here (i.e. Average Deviation, Skewness, Kurtosis, ..etc.). Note that there are two items that appear to be overlooked at this time and will appear in a future release of the OSLO on-line help system:

Cumulative Probability:

This function describes the probability (per cent) that the change in the error function (in the units of the error function itself) will be equal to the given value. The cumulative probability is given in both a plot and tabular form. Example: We will assume that the Monte Carlo analysis has produced 200 virtual systems and that the error function change for these 200 systems ranged from 0.1 to 1.6, with the median value being 0.85 (note: in this particular analysis, the original value of the error function is not important). We could say that the cumulative probability of the "error function change" would be 0% for an error function change of 0.1 and lower. Likewise, the cumulative probability of the "error function change" would be 100% for an error function change of 1.6 and higher. A 50% cumulative probability would correspond to an error function change of 0.85.

Relative Probability:

The relative probability summarizes the probability that the error function change will have a particular value. This function is plotted using blue lines comprising a bar chart. The longer the bar of the graph, the more likely a virtual system will be in close proximity to that particular value of the error function change.Example: We will assume that the Monte Carlo analysis has produced 200 virtual systems and that the error function change for these 200 systems ranged from 0.1 to 1.6. We will also assume that 5 of these virtual systems had an error function change with a close proximity to 1.1 and that there was only one virtual system with an error function change in close proximity to 1.55. In this case, the relative probability graph would show a bar that the 1.1 error function change that was 5x longer than the length of the bar at the 1.55 error function change.

Note that statistical analyses are only valid for a large number of samples. The smaller the number of virtual Monte Carlo systems that are generated, the less reliable the statistical analysis becomes.

Monte Carlo tolerance analysis is only available in OSLO Premium. "