The concept of a Monte Carlo simulation is commonly explained with the use of a coin toss analogy. Suppose you want to analyze the probability of a coin toss resulting in the coin landing with the presidential portrait facing up (heads) versus the other side facing up (tails). To perform a Monte Carlo simulation in an effort to gather these results, you perform a number of tests (or tosses) and record the results of each test.

If you toss the coin 5 times, you might find that 4 of the 5 tosses were heads and 1 of the 5 tosses was tails. Do these results, which are recorded from actual testing, offer accurate values for anticipating the probability of tossing heads?

If you extend your testing to include additional iterations (or tosses), your results change drastically. After 100 iterations, the probability of tossing heads or tails is likely to be very close to 50 percent. Therefore, the more iterations, the more accurate your probability value becomes.

Based on the coin toss analogy, it is clear that the number of iterations set for a Monte Carlo simulation is very important. While asking how many iterations to use is a common question, no set answer exists.

The time required to complete RBD calculations using simulation is significantly higher than the time required to complete an analytical calculation. As the number of iterations increases, the amount of computation time increases. Therefore, if your system is very large and very complex, you might attempt to calculate with fewer iterations in an effort to save time. For example, you might initially choose to start with 10,000 iterations. If, however, the time required to perform this calculation is too high, you might choose to decrease the number of iterations to save time.

With simulation, results between successive simulation runs on the same block diagram differ. To be able to duplicate results from one simulation run to the next, you must use the same random number seed, which is the numeric value from which to begin selecting random numbers for input. When the same random number seed is used for RBD calculations, assuming that no changes have been made to calculation properties, the results are the same from one simulation run to the next.