Capacity allocation under random slot assignment policy

Two non-communicating stations are considered that share a time-slotted communication channel on a random slot assignment basis. The optimal assignment policy is derived, and a simple strategy, based on a threshold test, is developed for the implementation of the optimal policy in a dynamically changing environment. Although the random slot assignment policy is inferior to the optimal periodic, fixed slot assignment policy, it is shown that it achieves the performance of that policy as the variance of the packet arrival process increases; the optimal periodic, fixed slot assignment policy is not, in general, implementable and it is very difficult for it to adapt to the conditions of a dynamically changing environment