Solving forward and inverse problems of the nonlinear Schrödinger equation with the generalized ${ \mathcal P }{ \mathcal T }$-symmetric Scarf-II potential via PINN deep learning
Jiaheng Li,Biao Li
Table 4. The correct NLSE with ${ \mathcal P }{ \mathcal T }$-symmetric Scarf-II potential and the identified twos obtained by learning V0 and W0, and the relative errors.
ItemNonlinear evolution equationRelative error [V0, W0](%)
PDE
Correct${\rm{i}}{\psi }_{t}=-{\psi }_{{xx}}-[2.91{{\rm{sech}} }^{2}(x)+0.3\mathrm{isech}(x)\tanh (x)]\psi +| \psi {| }^{2}\psi $[0, 0]
Identified (clean data)${\rm{i}}{\psi }_{t}=-{\psi }_{{xx}}-[2.91283{{\rm{sech}} }^{2}(x)+0.303\,28\mathrm{isech}(x)\tanh (x)]\psi +| \psi {| }^{2}\psi $[0.097, 1.093]
Identified (1 % noise)${\rm{i}}{\psi }_{t}=-{\psi }_{{xx}}-[2.89035{{\rm{sech}} }^{2}(x)+0.301\,65\mathrm{isech}(x)\tanh (x)]\psi +| \psi {| }^{2}\psi $[0.675, 0.552]