topic Poisson process for arrivals in FlexSim Forum

Poisson process for arrivals

maryamh11 — Thu, 04 Nov 2021 23:31:33 GMT

[ FlexSim 21.2.4 ]

Hi there,

If I use the Poisson process for arrivals with different volumes for AM/PM sessions as shown below (for AM session):

Object current = param(1);
treenode activity = param(2);
int rowNumber = param(3); //The arrival entry row number
int cycle = param(4); //The current cycle through the arrivals table (first iteration cycle = 1)
treenode processFlow = ownerobject(activity);
/***popup:StatisticalDistribution*/
return
/***tag:distribution*//**/poisson/**//**/(/**/
/***tag:par1*//**/33/**/
/***tag:par2*//**//**/
/***tag:par3*//**/, getstream(activity)/**/
/***tag:par4*//**//**/
/***tag:par5*//**//**/
/**/)/**/

I assume that 33 samples will be created and the arrivals depend on the Poisson probability. Is that correct?

And in this case, what would be the effect of using an "evenly spaced" spacing for arrivals?

Thanks!

Re: Poisson process for arrivals

Matthew_Gillespie — Fri, 05 Nov 2021 21:43:10 GMT

Putting a distribution in the Quantity column of the Date Time Source activity will just give you a variable number of arrivals. So, no, you probably won't get 33 arrivals. You'll get a number generated by the distribution. You'll generally get a number of arrivals around 33, but some replications you'll likely get less than 20 or more than 40:

Evenly Spaced will work the same way it would if you just put a number as the quantity. Say poisson(33) returns the number 20 in your model run and your arrival period is an hour, then 20 patients will be created during that hour spaced out every 3 minutes.

Re: Poisson process for arrivals

maryamh11 — Fri, 05 Nov 2021 21:49:22 GMT

Got it, thanks a lot!