Abstract
In wireless communications, the linear combination of lognormal random variables (LCLN) is a general model for the cochannel interference, which represents main limitation to radio link performance. Since the distribution function of LCLN is not known in the closed form and is difficult to compute numerically, simple and accurate approximations of the distribution are needed. In this paper we use two methods to approximate LCLN. One is Minimax approximation, which uses Minimax method to find a lognormal distribution approximating the distribution of LCLN. The other method also follows the rules of Minimax approximation, but uses a power lognormal distribution to improve the accuracy of approximation. Numerical examples are provided to compare the approximations we used here with the existing ones.