Traffic performance and mobility modeling of cellular communications with mixed platforms and highly variable mobilities

In modeling teletraffic performance of mobile cellular networks, previous work has made use of the concept of dwell time-a random variable that describes the amount of time a platform remains in a cell, sector, or microcell. The dwell time was characterized as a negative exponential variate or the sum of negative exponential variates. With a suitable state description, this allows the use of the memoryless property of negative exponential variates with the result that the problem of computing the traffic performance characteristics can be cast in an underlying framework based on multidimensional birth-death processes. Many alternative system configurations and issues can be investigated using this approach. In small cell (microcellular) systems, however, cell sizes tend to be much less regular in shape and size, and differences in paths traversed by mobiles have a large impact on dwell time realizations. Succinctly, some mobile platform classes have mobility characteristics that are highly variable, in that the dwell time standard deviation is greater than the mean. The previous models, in which the dwell time is a negative exponential or sum of exponential variates, may not be adequate in such cases because they can only accommodate dwell time variates for which the standard deviation does not exceed the mean. We extend the analytical framework so that highly variable mobilities can be considered while insights, tools, approaches, and formulations that are facilitated by the framework can be exploited. The approach allows computation of major teletraffic performance characteristics for cellular communications in which mobility issues are important. We consider multiple platform types as well as cutoff priority for handoffs. Additional issues can be considered using this same approach. Computational issues are discussed, and some theoretical performance characteristics are obtained to demonstrate the method and compare with previous work