Predicting evolutions of pulse characteristics along cavity position in passively mode-locked fiber laser via SSA-LSTM approach

Ultrashort pulse evolution along the cavity position is an extremely complicated computational process in passively mode-locked fiber laser systems. Herein, we present a data-driven model to predict the complex intracavity evolution of pulse characteristics along the cavity position using an integrated sparrow search al-gorithm (SSA) and long short-term memory (LSTM) approach instead of numerically calculating the nonlinear Schro center dot dinger equation (NLSE). Furthermore, the proposed model is compared with a separate LSTM network. The results predicted from the network agree well with those of the NLSE simulation. Moreover, the SSA-LSTM model exhibits excellent predictive performance.