6 views (last 30 days)
I am trying to find a probability distribution function (PDF) which can represent the availability of a certain component.
I get this availability data from my model. The model does a Monte-Carlo simulation and because I am using a random process to simulate failures on this component, I get an availability value from each run of the simulation.
Plotting the availability as a PDF makes possible to evaluate the confidence intervals of the output of my model.
The problem is that there's no such thing as availability greater than 100% or smaller than 0% and the distribution should take this into consideration.
I tried using the interactive tool called DFITTOOL to find a good fit, but non of them fulfill the rule previous described.
In the x-axis I want to show the availability values (0-100%) and on the y-axis, the probability of having a certain Availability. (The maximum of the distribution should be the mean Availability vector)
Any ideas are welcome!
Andrew Newell on 11 May 2011
A PDF doesn't have to be restricted to 100% (or 1, as I usually think of it) - as long as the integral of the PDF is 1, it's o.k. You may be thinking of a CDF, which should increase steadily from 0 to 1. So I'd go back to dfittool and fit the data to a CDF.