I would like to have your opinion or advices about how to represent arrivals when you simulate an ambulatory clinic where all patients have an appointment.
What would be the best practice:
1) Use the historical data (if there are available) from the real appointment system and insert them into the appointment table. So it means that you will test your scenario only on one historical period and you will make few replications of a long time model for your experimentations.
2) Represent the different arrival patterns based on distribution and probability. It means that you define with your historical data the 2 probabilty (number and patient type) for all your time slot. And then insert them into the custom arrival table.
For example, if you want to represent your arrival on Monday at 08:00 with 2 possible patient types : 01:08:00 – 01:08:15 P(50,0,25,1,25,2) P(50,1,50,2)
In this case you will only replicate one week few time for the experimentation.
What is your opinion on this? Or, how would you represent arrivals into an outpatient clinic model to fit with reality ?
Thanks in advance.