Hello

How are RAM simulation studies done via a) Monte Carlo and b) Discrete event approcahes - different?

Given that they are and this can be explained relatively simply; then how would we know when to prefer one method versus the other?

Regards
Wayne

Wayne,

Monte-Carlo simulation is used whenever analytical models can’t cope with the complexity of most situations in practice due to the interaction of various variables. With regard to reliability and maintainability issues, Monte-Carlo is advantageous to model discrete events (the moment a random failure takes place, the moment a repair task starts, the moment a spare part becomes available, and so forth) such as in the case of parallels or series of items (or still, complex arrangements – the ones that incorporate items which cannot be said to be either in parallel or series) that interact as a consequence of a failure originated in any one of them. In this particular case, analytical models are definitively unable to model the interactions. There are many other instances where Monte-Carlo simulation proves to be more advantageous despite analytical models do exist. This is the case of, again, parallel, series or complex arrangements, where the aim is often to calculate the system availability. Analytical expressions are sometimes only approximative. On the contrary, with Monte-Carlo, you use logic to create your model (something that comes naturally out of your head and you understand fully) as opposed to the analytical expressions. These were deduced and written down in books sometimes considering assumptions that don’t hold in practice in many occasions. Monte-Carlo modelling allows you much more freedom in the design of your system and facilitates the appraisal of its performance.

I use both methods in my analysis whenever analytical expressions exist, just for the sake of validation. The results become consequently much more trustful.

To know more and experiment some Monte-Carlo models, the Barringer website http://www.barringer1.com is quite informative.

Regards,

Rui

This message has been edited. Last edited by: <Rui Assis>,