An approximation and its applications in wireless networks performance analysis: Research Articles

In the literature, there are two common assumptions for the tele-traffic parameter in analyzing the wireless network performance, that is, the tele-parameter follows a specific probability density function (pdf) and additionally the pdf exists closed-form Laplace Transform (LT). However, taking into account the cell irregular shape, the specific pdf may be unavailable while only the measured statistical moments are available. Moreover, the pdf function may not exist a closed-form LT, for example, lognormal distribution function. In this paper, based on the Central Limit Theorem and hyper-Erlang universal approximation property, we propose an approximation method applicable in the situations when only the statistical moments are available or LT of pdf does not exist. We then employ the technique in diverse applications, including the performance analysis of wireless network and the cost evaluation of mobility management. Extensive numerical examples demonstrate the good approximation capability to the exact formula and the simulation results. Copyright © 2006 John Wiley & Sons, Ltd.