I have a problem where i'm trying to characterize a system that has several independent variables. Each of these variables has an associated variance. What i'm doing is running a system model with a monte carlo analysis and trying to determine the overall variance on a system variable. I know the variances on all of the independent variables and for the most part they have a normal distribution. However, a few of them have a skewed distribution. For example, say the mean for one variable is 80 and the plus 3-sigma is 84 but the minus 3-sigma is 70. What kind of distribution do i use to model this? I looked at the log-normal distribution, but i'm having difficulty applying this to my problem. I hope this makes sense.
Does anyone have any suggestions?
Does anyone have any suggestions?