A TEST FOR THE DETECTION OF A POISSON PROCESS

A statistical test is presented to detect the presence of a Poisson input source into a queueing system. The fact that the arrival times of a Poisson input source are distributed independently and uniformly over time is used in conjunction with the central limit theorem to develop a test, based on the sum of the arrival times over a given interval of time, for the hypothesis that an input source into a queueing system is Poisson. Failure to reject this hypothesis gives one an extra degree of assurance that traditional queueing models assuming a Poisson input are applicable.