I am looking for hints on what type of literature to consult to learn to do this.

Quote:

A proposed project has a projected four-year life. (The projected life of a project is typically not known with certainty, and generally should be treated as the realized value of a random variable, but here we treat it as known.) Construction will take place during year 1. The construction costs are estimated to be normally distributed, with mean $4,000,000 and standard deviation $500,000. Revenue begins in year 2, and is estimated to be normally distributed with mean $14,000,000 and standard deviation $1,500,000. Revenue in years 3 and 4 is estimated to be normally distributed, with mean equal to the preceding year’s sampled revenue, and standard deviation set at 10% of the preceding year’s sampled revenue. (This year-to-year relationship reflects a dependency among input variables. More generally, correlated input variables can be described in spreadsheet add-ins such as @RISK by providing matrices of correlation coefficients in models.)

Operating expenses in years 2, 3, and 4 are estimated to range from 60% to 80% (uniformly distributed) of the corresponding year’s revenues. Taxes are 40% of the post-expense revenues).

Central to simulations is deciding what distribution to use. How do you decide what distribution makes sense if you have no prior research? For example say I wanted to simulate future usage of a service. How do you go about deciding what distribution it makes sense to sample from (I am guessing this will be poisson in this case, but in a lot of other cases it won't be obvious). :) ]]>

I am an internship student and for my company I have to develop a method to test if their material is brittle or not. They bought a new machine so first I have to start with an MSA (measurement system analysis).

I would like to test the repeatability and reproducibility (Gage R&R). The problems are : I can only use 1 operator (myself), the samples are destructive (they burst during testing so the same sample can't be tested a 2nd time), I have no reference values since this is the first time this machine is used with this kind of material. I try to avoid the part to part variation by using thick paper as reference, this paper is almost uniform in thickness and in bursting strength. So for now I have the following data:

50 samples of thick paper tested once, all tested by the same operator, 50 single results...

I tried to use the nested Gage R&R in minitab but I keep getting errors because it requires 2 operators and it requires replicate results (although the troubleshooting says it's not necessary and that this method is perfect for destructive testing with 1 operator)

Can someone please help me?

Thank you very much in advance ! ]]>