Hi Zahid,

1. When the MCS has finished the computations, the best to do is to construct the pdf and cdf and after that you can find which distribution gives the best approach to the graph, this way you could use another less time consuming method (such as FORM). But myself I would choose MCS to FORM or other reliability analysis methods.

AND always use a method to evaluate the number of needed iterations in the MCS!

2. Before the simulation, you can use variance reduction techniques, such as Latin Hypercube, in order to reduce the calculation time. But take into account that this approach is no more a "pure" stochastic process...

Regards

It is good to use variance reduction techniques. Like energy cut off and forced collision

One of the methods to reduce the scattering of Monte Carlo data is numerical smoothing. For details, see https://en.wikipedia.org/wiki/Smoothing.