Fractional physics-informed neural networks for necklace beams with angular momentum in saturable nonlinear media under an annular trapping potential
Phys. Rev. A 114, 013525 – Published 30 July, 2026
DOI: https://doi.org/10.1103/j4q3-ylw1
Abstract
We propose a fractional physics-informed neural network (fPINN) framework to investigate the propagation dynamics of necklace beams trapped in an annular potential within a saturable nonlinear medium. The interplay among fractional-order diffraction, potential confinement, and nonlinearity gives rise to robust necklace beams carrying an arbitrary fractional orbital angular momentum. Using the fPINN approach, we systematically explore the propagation dynamics of such beams. While the annular potential effectively suppresses radial expansion, the combination of fractional diffraction and saturable nonlinearity significantly enhances the beam's robustness. Our findings unveil physical behaviors in fractional nonlinear optical systems and highlight the effectiveness of fPINNs for solving complex fractional-order nonlinear problems.
