An Improved Perturbation Theory for Optical Solitons Near the Zero-Dispersion Point

In communication lines operated near the zero-dispersion point, the third-order dispersion (TOD) can strongly affect the pulse dynamics. Hence, it is necessary to develop proper theoretical tools for this case. Besides the numerical methods, a perturbation theory that can take into account the bandwidth-limited amplification (BLA) and nonlinear gain (NLG) or losses is required. The latter perturbations can considerably reduce soliton interactions. With regard to a combined effect of TOD and BLA, it was shown [1] that, e.g., a decay of a two-soliton bound state caused by TOD can be prevented by BLA. A new idea put forward in [2, 3] was that BLA which compensates for the linear and nonlinear losses can, simultaneously, absorb the resonance radiation emitted under the action of TOD, thus lending the soliton a much better stability. This effect has been experimentally observed in [4]. For large values of TOD, however, the soliton was obtained numerically [3] in a state with a conspicuous velocity, which differed from that predicted by the usual perturbation approach [5] by a factor of 2.