Approximating a Sum of Random Variables with a Lognormal

A simple, novel, and general method is presented in this paper for approximating the sum of independent or arbitrarily correlated lognormal random variables (RV) by a single lognormal RV. The method is also shown to be applicable for approximating the sum of lognormal-Rice and Suzuki RVs by a single lognormal RV. A sum consisting of a mixture of the above distributions can also be easily handled. The method uses the moment generating function (MGF) as a tool in the approximation and does so without the extremely precise numerical computations at a large number of points that were required by the previously proposed methods in the literature. Unlike popular approximation methods such as the Fenton-Wilkinson method and the Schwartz-Yeh method, which have their own respective short-comings, the proposed method provides the parametric flexibility to accurately approximate different portions of the lognormal sum distribution. The accuracy of the method is measured both visually, as has been done in the literature, as well as quantitatively, using curve-fitting metrics. An upper bound on the sensitivity of the method is also provided.

[1]  A. Safak,et al.  Statistical analysis of the power sum of multiple correlated log-normal components , 1993 .

[2]  John G. Proakis,et al.  Probability, random variables and stochastic processes , 1985, IEEE Trans. Acoust. Speech Signal Process..

[3]  D. Owen.  Handbook of Mathematical Functions with Formulas , 1965 .

[4]  Bruno O. Shubert,et al.  Random variables and stochastic processes , 1979 .

[5]  Abbas Jamalipour,et al.  Wireless communications , 2005, GLOBECOM '05. IEEE Global Telecommunications Conference, 2005..

[6]  Fortunato Santucci,et al.  On SIR fade statistics in Rayleigh-lognormal channels , 2002, 2002 IEEE International Conference on Communications. Conference Proceedings. ICC 2002 (Cat. No.02CH37333).

[7]  A. Goldsmith,et al.  Area spectral efficiency of cellular mobile radio systems , 1997, 1997 IEEE 47th Vehicular Technology Conference. Technology in Motion.

[8]  Norman C. Beaulieu,et al.  An optimal lognormal approximation to lognormal sum distributions , 2004, IEEE Transactions on Vehicular Technology.

[9]  Norman C. Beaulieu,et al.  Estimating the distribution of a sum of independent lognormal random variables , 1995, IEEE Trans. Commun..

[10]  T. T. Ha,et al.  On the sum of multiplicative chi-square-lognormal random variables , 2001, GLOBECOM'01. IEEE Global Telecommunications Conference (Cat. No.01CH37270).

[11]  J. Couper.  Sensitivity and Uncertainty Analysis , 2003 .

[12]  D. C. Schleher,et al.  Generalized Gram-Charlier series with application to the sum of log-normal variates (Corresp.) , 1977, IEEE Trans. Inf. Theory.

[13]  William A. Janos,et al.  Tail of the distribution of sums of log-normal variates , 1970, IEEE Trans. Inf. Theory.

[14]  Slimane Ben Slimane,et al.  Bounds on the distribution of a sum of independent lognormal random variables , 2001, IEEE Trans. Commun..

[15]  M. D. Austin,et al.  Exact cochannel interference analysis for log-normal shadowed Rician fading channels , 1994 .

[16]  A. Annamalai,et al.  An unified numerical approach for computing the outage probability for mobile radio systems , 1999, IEEE Communications Letters.

[17]  Theodore S. Rappaport,et al.  Statistical analysis of co-channel interference in wireless communications systems , 2001, Wirel. Commun. Mob. Comput..

[18]  A. Abu-Dayya,et al.  Outage probabilities in the presence of correlated lognormal interferers , 1994 .

[19]  Milton Abramowitz,et al.  Handbook of Mathematical Functions with Formulas, Graphs, and Mathematical Tables , 1964 .

[20]  C. Ho.  Calculating the mean and variance of power sums with two log-normal components , 1995 .

[21]  Hirofumi Suzwi,et al.  A Statistical Model for Urban Radio Propagation , 1977 .

[22]  H. Suzuki,et al.  A Statistical Model for Urban Radio Propogation , 1977, IEEE Trans. Commun..

[23]  Jean-Paul M. G. Linnartz,et al.  Exact analysis of the outage probability in multiple-user mobile radio , 1992, IEEE Trans. Commun..

[24]  R. Barakat.  Sums of independent lognormally distributed random variables , 1976 .

[25]  Huaping Liu,et al.  Error performance of a pulse amplitude and position modulated ultra-wideband system over lognormal fading channels , 2003, IEEE Communications Letters.

[26]  Gordon L. Stuber,et al.  Principles of Mobile Communication , 1996 .

[27]  Ainslie,et al.  CORRELATION MODEL FOR SHADOW FADING IN MOBILE RADIO SYSTEMS , 2004 .

[28]  H. R. Anderson.  Signal-to-interference ratio statistics for AM broadcast groundwave and skywave signals in the presence of multiple skywave interferers , 1988 .

[29]  J. Naus.  The Distribution of the Logarithm of the Sum of Two Log-Normal Variates , 1969 .

[30]  S. Schwartz,et al.  On the distribution function and moments of power sums with log-normal components , 1982, The Bell System Technical Journal.

[31]  Fredrik Berggren,et al.  A simple bound on the outage probability with lognormally distributed interferers , 2004, IEEE Communications Letters.

[32]  D. Cacuci,et al.  SENSITIVITY and UNCERTAINTY ANALYSIS , 2003 .

[33]  F. Graziosi,et al.  Second order statistics of the SIR for cellular mobile networks in the presence of correlated co-channel interferers , 2001, IEEE VTS 53rd Vehicular Technology Conference, Spring 2001. Proceedings (Cat. No.01CH37202).

[34]  L. Fenton.  The Sum of Log-Normal Probability Distributions in Scatter Transmission Systems , 1960 .

[35]  Norman C. Beaulieu,et al.  Comparison of methods of computing correlated lognormal sum distributions and outages for digital wireless applications , 1994, Proceedings of IEEE Vehicular Technology Conference (VTC).