Jump to content

Data incorporation


Mthost
 Share

Recommended Posts

Hello :D

I am currently sitting with at dataset of incomming calls for a call center. But we cannot get the right distribution for the incoming calls, that somewhat matches our data. We have tried to make Random.Poisson, but the result is not close to the real numbers in the data set. Our method so far has been to use statfit 3 to calculate a discrete distribution for the incoming calls interarrival time, but this does still not seem to work.


Moreover, the data seem to show no correlation between waiting time, inbound time, duration and the date.


TLDR: We do not know how to convert the data shown into a distribution for the source. Moreover, the distribution of calls throughout each day differ significantly.


Any tips would be very helpfull :)

Edited by Guest
Link to comment
Share on other sites

One option would be to have an arrival table, instead of fitting the data to a distribution - the arrival table would include the time that the arrival occurs (based on your table above) and then you would reference the arrival table from the Source using Arrival Mode == 'Arrival Table'. With an arrival table, you could also include an Arrival Time Deviation (Source - under Other Arrival Stream Options) property as a distribution, which would add some randomness to the arrival time.


Another option is to use the Arrival Mode == 'Time Varying Arrival Rate' where you would specify a Rate Table (Data tab) that would include varying rates for different times of day. The interval size and number of intervals can be easily changed. The Rate (events per time period) would include the mean of an exponential distribution and would allow the rates to change hourly or every x hours.


You may wish to review a few of the Sample SimBit Solutions (Support ribbon) under the Arrival Logic category of SimBits for more options.

Link to comment
Share on other sites

×
×
  • Create New...