Abstract-Effect of birefringent disorder on the Bit Error Rate (BER) in an optical fiber telecommunication system subject to amplifier noise may lead to extreme outages, related to anomalously large values of BER. We analyze the Probability Distribution Function (PDF) of BER for various strategies of Polarization Mode Dispersion (PMD) compensation. A compensation method is proposed that is capable of more efficient extreme outages suppression, which leads to substantial improvement of the fiber system performance. [7]. Optical noise generated in optical amplifiers represents another impairment that may not be reduced/compensated and, therefore, should be also considered in any evaluation of a fiber system performance [8]. BER calculated for a given realization of birefringent disorder by means of averaging over the amplifier noise statistics constitutes an appropriate object to characterize joint effect of the two impairments. In two preceding papers [9], [10] we have demonstrated that the probability of extreme outages (values of BER much higher than typical) is substantially larger than one could expect from naive Gaussian estimates singling out effects of either of the two impairments. The natural object of interest is the PDF of BER and, specifically, the PDF tail corresponding to anomalously large BER. In [9] we have developed a consistent theoretical approach to calculating this tail. The case when no compensation is applied and also the effect of the simplest "setting the clock" compensation on the PDF tail suppression have been discussed in [9]. Then our investigation was extended to study effects of the standard first-and higherorder compensations on extreme outages [10]. In the present letter we propose a compensation scheme that appears to be more efficient in reducing the extreme outages compared to the traditional high-order compensation scheme with the same number of the compensating degrees of freedom.
Index Terms-We consider the return-to-zero (RZ) modulation format, when optical pulses are well separated in time t, and thus can be analyzed as individual objects. We represent the pulse intensity measured at the system output aswhere G(t) is a convolution of the electrical (current) filter function with the sampling window function. The twocomponent complex field Ψ (t) describes the output optical signal (the components correspond to two polarization states of the signal). The linear operator K in Eq. (1) represents optical filtering, it may also account for a compensating device. The compensating part of the linear operator, K c , is applied first, i.e. before filtering described by K f , resulting inIdeally, I takes two distinct values depending on whether the information slot is vacant or filled. However, the impairments enforce deviations of I from those fixed values. If the output signal intensity exceeds the decision level I d , then "1" is associated with the slot, otherwise the slot is labeled by "0". Sometimes the information is lost, i.e. the initial "1" is detected as "0" at the output...