I require reading random process times from global table

Question

question

Jaime G asked Apr 27, '20 tannerp commented May 4, '20

I require reading random process times from global table

Dear collaborators. I must read random process times from a table whose cells I have declared in flexscript format. But simulating the model warrants this error message:

time: 16.289937 exception: FlexScript exception: Property "dataType" accessed on invalid node . at <no path>

Annex the model, which includes the global table: Time Process invoked from each processor.

Modelo: Flujo empleando tiempos aleatorios desde tabla.fsm

Software Version:

FlexSim 20.0.0

flexsim 20.0.0

flujo-empleando-tiempos-aleatorios-desde-tabla.fsm (26.3 KiB)

Answer 1 · 2020-04-27T15:43:38Z

Regan Blackett answered Apr 27, '20 Jaime G commented Apr 27, '20

While you can use the expression:

getstream(current)

within the cell of a global table, it's a little difficult, because there is no built in reference to 'current' in the code for a table cell; you would have to construct one yourself.

What is easier is just using a stream number parameter instead so use something like:

normal(12,1.2,1);
return triangular(5,12,1.2,2);
return uniform(3,5.6,3);

with numbers in the last parameter in the distribution instead of getstream()

· 1

Jaime G commented · Apr 27, 2020 at 08:04 PM

@Regan Blackett

Dear Regan, thank you very much for your collaboration.

0 ·

Answer 2 · 2020-04-27T16:33:37Z

Jeff Nordgren answered Apr 27, '20

@James G,

Attached is your model with the changes that I've made. @Regan Blackett answer is correct. This just shows you how to implement it into your model.

If you have any questions or problem, please let us know.

Thanks.

flujo-empleando-tiempos-aleatorios-desde-tabla-JN1.fsm

flujo-empleando-tiempos-aleatorios-desde-tabla-jn1.fsm (32.8 KiB)

Answer 3 · 2020-04-27T22:39:24Z

Jaime G answered Apr 27, '20 Joerg Vogel edited Apr 28, '20

Dear collaborators . In order to generate seed randomly, I tried this code in the global table and it gave no error. From a statistical point of view, is this correct?

double Semilla = duniform(1,100,4);

return normal(12,1.2,Semilla);

· 1

Joerg Vogel commented · Apr 28, 2020 at 02:24 AM

No, it is not correct, because you load your model and the seed values are the same again. But if you have a license and you set a global preference on the environment tab to “initialize random streams based on system time“ you get around this problem. You don’t even need a duniform to choose a different stream anymore, because you choose a first value somewhere else on the stream then before.

But be aware of the main menu Statistics option “Repeat Random Streams“. If it is set for finding errors in your model, every run gets same results, This behavior is default always set to on in express licensed version and you can‘t get around it.

0 ·

question