An enhanced inheritance algorithm for optimization of multiwavelength fiber Raman amplifier

Abstract

Multi-Wavelength Fiber Raman Amplifier (MW-FRA) has become an essential component in DWDM system due to its broad bandwidth, flexibility and excellent performance. This paper is first dedicated to an accurate modeling of the MW-FRA for a propagation equation, which includes fiber loss, Stimulated Raman Scattering (SRS), Rayleigh backscattering, spontaneous emission noise and thermal noise. Then based on the reasonably simplified model it focuses on a numerical calculation with initial pump wavelength and power for the acquisition of on-off gain and flatness of signals, which are taken as feedback information to adjust the corresponding pumps. An inheritance algorithm is applied in the process of such iteration, which moves on until the target gain and flatness are achieved. According to the Raman gain spectrum characteristics, we can improve the algorithm to speed up the calculation and to make it more practically applicable. The design idea here is implemented by using C language. With input target gain and flatness, initial value of signals and pumps, the program can output the optimized results of pump wavelength and power. Besides, the simulation results from employing VPI are demonstrated, which proves the algorithm is effective.