Genetic Algorithm-Driven Optimization of Multi-Tone Phase Modulation Signals for SBS Suppression in High-Power Fiber Amplifiers

Abstract

To meet the demands of high-power, narrow-linewidth fiber lasers in defense and industrial applications, stimulated Brillouin scattering (SBS) has emerged as a critical bottleneck limiting power scaling. Conventional SBS suppression techniques, such as white noise and pseudo-random binary sequence (PRBS) modulation, often suffer from uncontrollable spectral broadening and limited tunability of the spectral shape. In this work, we propose a novel drive signal design methodology that leverages genetic algorithm (GA)-based optimization of frequency-domain amplitude distributions within a multi-tone modulation framework. This approach circumvents the randomness introduced by traditional binary modulation and enhances both the controllability and determinism of the signal. A simulation model based on the standard coupled three-wave and rate equations is utilized to systematically evaluate the impact of the optimized spectral profiles on SBS threshold enhancement. Simulation results demonstrate that, under a fixed 11 GHz of 20 dB RMS spectral bandwidth, the proposed scheme yields an SBS threshold improvement of approximately 400 W and 180 W over Gaussian-shaped and flat-top spectra, respectively. Experimental validation on a 2-kW-class fiber amplifier platform confirms these trends, with measured enhancements of 350 W and 120 W. These findings substantiate the feasibility and practical relevance of the proposed method, offering a new pathway for spectral engineering and SBS mitigation in next-generation high-power fiber laser systems.