Reinforcement Learning for Efficient Identification of Soliton System Parameters Across Expansive Domains

Optical solitons in mode-locked fiber lasers and optical communication links have various applications. Thestudy of transmission modes of optical solitons necessitates the investigation of the relationship between theequation parameters and soliton evolution employing deep learning techniques. However, the existing identificationmodels exhibit a limited parameter domain search range and are significantly influenced by initialization.Consequently, they often result in divergence toward incorrect parameter values. This study harnessed reinforcementlearning to revamp the iterative process of the parameter identification model. By developing a two-stageoptimization strategy, the model could conduct an accurate parameter search across arbitrary domains. Theinvestigation involved several experiments on various standard and higher-order equations, illustrating that theinnovative model overcame the impact of initialization on the parameter search, and the identified parametersare guided toward their correct values. The enhanced model markedly improves the experimental efficiency andholds significant promise for advancing the research of soliton propagation dynamics and addressing intricatescenarios.