Adding or subtracting process time from a fitted distribution as a parameter

[jgolob]

[ FlexSim 25.0.0 ]

My model contains a fitted distribution of environmental service clean times of operating rooms based on real data. I am looking to add or subtract minutes from random clean times selected from this distribution as a model parameter. For example, if the randomly selected clean process time for a room was 10.5 minutes, I want to set a parameter to -2 which would subtract 2 minutes from the randomly selected time and yield 8.5 minutes. This would allow me to see how the other performance measures respond under a the gross assumption of 2 minute decrease in clean times. This fitted distribution is currently called each time one of the 20 operating rooms is cleaned.

I have reviewed the following answers from this forum:

https://answers.flexsim.com/questions/109886/how-to-reference-to-a-statistical-distribution-in.html

https://answers.flexsim.com/questions/95419/how-to-change-the-empirical-distribution-in-experi.html

The picture below shows my latest attempt. Although no errors, there was no change to my output. Do I need to put something in the 'On set' property? I realize I am sacrificing true randomness, but this is the best way I could think how to study impacts on changes in clean times. Is what I am looking to do possible?

[moehlmann_fe · Accepted solution]

The parameters should work as you set them up. Though you should consider using the Math.max() function to prevent negative values from being generated.

Can you show where/how you call the parameter in the model?

Linking to the empirical distribution doesn't do anything on its own. It only makes that reference available as a variable in the On Set code. But I don't see why that would be necessary here.

[jgolob]

Thanks Felix. I was under the impression when I referenced the empirical distribution 'CleanUpTimePerCase' within the parameter 'NewCleanTime' a link was automatically created.

Currently I am calling the empirical distribution 'CleanUpTimePerCase' in multiple process activities as per screen shot below. How would I call the new parameter 'NewCleanTime'?

Also, where and how would I utilize the Math.max() function. I do want to avoid negative values from being generated as you mentioned.

[moehlmann_fe]

By using "Empirical()" you get a value based on the empirical data. Not the fitted distribution and not the parameter. You need to use "Model.parameters.NewCleanTime" instead.

The expression of that parameter should then be:

Math.max(Model.parameters.CurrentCleanTime + Model.parameters.ChangeCleanTime, 0)

It will return either the sum of the two other parameters or 0, whichever is higher. (You can of course also use a different number than 0 for the second value.)

[jgolob]

Thank you Felix. The model is working a required.