In a tweet https://twitter.com/elonmusk/status/678679083782377472 Mr Musk states that they are using Monte Carlo analysis to determine mission success.

My Question(s):

  • how does that work exactly?
  • what parameters flow into this simulation?
  • is this standard in the space industry?

And now the really important part: I'm in software engineering and I'd like to know if and how this is applicable to software engineering (project success, failure rates, etc) too? Where do i start if I would like to do this?

  • 1
    $\begingroup$ To get you started: en.wikipedia.org/wiki/Monte_Carlo_method $\endgroup$ – Hobbes Dec 22 '15 at 11:09
  • $\begingroup$ I suspect, from the context, that the analysis being described was specifically limited to the landing success, so on that basis the specific analysis isn't a standard thing. Beyond that its wide open, monte carlo analysis is used in all kinds of ways and any launch might be supported by hundreds (thousands) of supporting bits of analysis. If you really want the all-up prediction of mission success then there are relatively few approaches, all high-level and none of which work really well in my opinion. There wouldn't be an insurance business otherwise, but its a good question all the same. $\endgroup$ – Puffin Dec 22 '15 at 13:08
  • $\begingroup$ There is discipline called software reliability. I think it relates to counting the progression of errors found and fixed during development. That's as much as I know of it and I don't know how widely its practised. $\endgroup$ – Puffin Dec 22 '15 at 13:12
  • 1
    $\begingroup$ I've already read the article on wikipedia and a few others. The question is how does space x (and others) use this for forecasts of mission success. If they rely on it in this way, it must be important and there must be some reasoning behind the decision to use this method! $\endgroup$ – mike Dec 22 '15 at 13:39
  • $\begingroup$ I assume they've analysed every piece of data for every recorded rocket launch, and tried to determine what effects E.G. weather, humidity, temperature and who knows what else have had on the success of aspects of previous launches. And what effect those aspects have on the likely success of their own launches. Plus, they have over 30 of their own launches to analyse, test simulations on etc. including 3 previous landing-attempts. $\endgroup$ – Kaz Dec 22 '15 at 13:59