A Deep Reinforcement Learning Algorithm for Smart Control of Hysteresis Phenomena in a Mode-Locked Fiber Laser

We experimentally demonstrate the application of a double deep Q-learning network algorithm (DDQN) for design self-starting fiber mode-locked laser. In contrast to static optimization system design, DDQN reinforcement is capable learning strategy dynamic adjustment cavity parameters. Here, we apply stable soliton generation in laser exploiting nonlinear polarization evolution mechanism. The learns hysteresis phenomena that manifest themselves as different pumping-power thresholds regimes diverse trajectories adjusting optical pumping.