For Monte Carlo AHP, I used 1000 simulations. However, to make sensitive analysis, one may use many simulations for more accurate and enhanced results.
Thanks Evgeny. It's interesting, any reason/literature to support 1024, neigther 1000 nor 1100. I have around 100 parameters, so, it means I need way more than 1000 simulations?
Thank you Stefan for the wonderful insight. Overall purpose is to see how the variability of diff. input parameters affect an output parameter, which are the most sensitive inputs? or for screening purposes. Yes, they are highly non-linear between input and output. There might be correlation among some input parameters. Inputs are around 100! and need to explore tail region too. Any suggestive simulations or a good paper/s on these? Thanks again.