Don't understand the values of the counter

[TristanFAURE]

Hello all
thank you for providing this library it is quite helpful

I'm trying to simulate a patient flow based on the tutorial "Plotting from resource monitors"

You can find bellow the code but basically I have several (around 50 by month) patients by month, one queue, and patients stayed for an average period of 72h in one bed and I want to simulate several scenario changing the numberOfBeds or TimeInHopital vars

If i perform the simulation with 15 beds the curves seems OK and I see how beds (resources) are used :

If I performe the simulation with 5 beds I see that resource usage explodes :

I would expect to have my resource usage limited by my number of beds (15 or 5), but maybe there is something I did not get from DES theory

Thank you in advance !

library(simmer)
library(simmer.plot)

timeInHopital <- 3 * 24 # mean, hours
numberOfBeds <- 5
nbMonth <- 12
numberOfPatients <- 50 * nbMonth
nbHours = nbMonth * 31 * 24

set.seed(1234)

bank <- simmer()

customer <-
  trajectory("Customer's path") %>%
  set_attribute("start_time", function() {now(bank)}) %>%
  seize("counter") %>%
  timeout(function() {rexp(1, 1/timeInHopital)}) %>%
  release("counter")

bank <-
  simmer("bank") %>%
  add_resource(name = "counter",capacity = numberOfBeds) %>%
  add_generator("Customer", customer, function() {c(0, rexp(numberOfPatients, 1/10), -1)})

bank %>% run(until = nbHours)
#> simmer environment: bank | now: 248.706140163915 | next: 
#> { Monitor: in memory }
#> { Resource: counter | monitored: TRUE | server status: 0(1) | queue status: 0(Inf) }
#> { Source: Customer | monitored: 1 | n_generated: 20 }

plot(bank,
     what = "resources",
     metric = "usage",
     names = "counter",
     items = "system",
     steps = TRUE)

[Enchufa2]

The add_resource function provides an infinite queue by default. The capacity is the number of beds you defined (the number of "servers"), and, when you plot "system", you see server + queue, so you see your patients accumulate in the queue. If you want your patients to be dropped from the simulation, you must limit the queue, or even remove it completely (setting queue_size=0).

[TristanFAURE]

Got it, it is clearer ! Thank you for the quick answer
I will try to do some changes on queue_size