Improving call admission control procedures by using hand-off rate information

This paper introduces a general decision model, in the shape of a Markov Decision Process, as an instrument to analytically compare the behavior of call admission control policies. This approach allows the study of a wide class of policies, including well-known pure stationary as well as randomized policies, in a way that explicitly incorporates the dependency between the hand-off rate and the system state, assuming that the hand-off rate arriving to a cell is proportional to the occupancy level of the adjacent cells. In particular, some well-known non-preemptive prioritization schemes are analyzed, including the Cutoff Priority Policy (CPP), which consists of reserving a number of channels for the high priority requests stream. Using our analytical approach, we prove the optimality of CPP within the analyzed class.