On the identification of a distribution for the arrivals to a queueing system: an application from bank data

Queues are an integral part of life. The literature available concentrates on Poisson arrival and exponential service time. The present paper focuses on statistical inference on the parameters of queueing models. The work is data driven. The dataset considered is the queueing system in banks which is publicly available. The number of servers has been estimated for a steady state system. The data has been tested for normality. The goodness of fit test has been done for the arrival pattern, and it is found that the discrete analog of smallest extreme value distribution is a good fit to the data which is an alternative to the widely used Poisson distribution.