abeham
commented
4 years ago Difficult to say, can you be more specific in what you think is wrong? The presence of a single value of 8.6 seconds is not really an indication that something is wrong. What makes you think 8.6s would be an outlier? There is a 1.6% chance to obtain a value that is at most 8.6 if you sample from N(30, 10). With 240 samples as you have in your data, there is a 97.9% chance to obtain at least one value that is 8.6 or less.
Here is the density estimation of your samples overlaid with 5 density estimations from actual normal distributions. Note that your peak is shifted to the right, because NormalPositive is implemented using rejection sampling and thus resampling until a positive value is obtained. As you see below, a normal distribution would even return negative values given these distribution parameters.
AltL90
I don't thinks (TimeoutNormalPositive) works correctly. It has two input variables TimeMu and TimeSigma. The way I understood it is ==> TimeOut = TimeMu +- TimeSigma If so why my program delivers something else?
My program is just the simple shop example, but without the packer. It shows how much time would it take the machine to produce all of the products.
In the Program TimeMu = 30 s and TimeSigma = 10 s Take a look at result Line 18 (Machine production time: 00:00:08.6190997)
program program.txt results res.txt