A statistical design approach for gigabit-rate fiber-optic transmission systems

A statistical design approach is proposed for the design of fiber-optic transmission systems (FOTSs). It yields increased regenerator spacings up to about 50% over those provided by the traditional worst-case design. The method involves assuming most of the system parameters to be random variables with known probability distributions. The probability distribution of the regenerator spacing is then numerically found by a Monte Carlo simulation from which the regenerator spacing with a given success probability can be determined. A thorough statistical treatment of attenuation, dispersion, and chirp limits is presented, and the effects of varying the probability density functions of underlying optical system parameters are examined. The numerical results shown demonstrate the repeatability of the proposed statistical design approach.<
>