-
The inverse problem of reconstructing the PT-symmetric potential in a class of (1+1)-dimensional nonlinear Schrödinger equation is investigated in this work. The governing equation is given as follows: $ \text{i}u_t(x,t) + u_{xx}(x,t) + \alpha\left| u(x,t) \right|^2 u(x,t) + \beta\left| u(x,t) \right|^4 u(x,t) + V_{\rm PT}(x) u(x,t) = 0, $where u(x, t) denotes the wave function in dimensionless coordinates, and the PT-symmetric potential VPT(x) = V(x)+iW(x) consists of a real part V(x) and an imaginary part iW(x), satisfying the symmetry conditions V(x) = V(–x) and W(x) = –W(x). In this inverse problem, partial boundary values of the wave function are known, while the potential $ V_{\rm PT}(x) $ is the unknown to be reconstructed. To address this challenge, we construct a three-level finite difference scheme for the corresponding forward problem, discretizing both the wave function and the potential. This approach leads to a nonlinear system of equations that links the known wave data to the unknown potential values. To simplify the computation, we separate the real and imaginary parts of this system and reformulate it as a real-valued nonlinear system of equations. For the numerical solution, we employ an inexact Newton method to iteratively solve the resulting nonlinear system. In each iteration, the Jacobian matrix is approximated numerically. To ensure that the reconstructed potential strictly satisfies the PT-symmetry, a parity correction mechanism is introduced at the end of the iteration process. We conduct numerical experiments under both noise-free (exact data) and noisy (inexact data) conditions. The results indicate that in both cases, the proposed method converges within a limited number of iterations and keeps the reconstruction error within the order of 10–3. These findings verify the effectiveness and robustness of this method in solving inverse problems of PT-symmetric potentials, and provide a new idea and powerful method for related numerical applications. -
Keywords:
- PT-symmetry /
- Schrödinger equation /
- potential function /
- inverse problem
[1] Liu Y P, Gao Y T, Wei G M 2012 Phys. A Stat. Mech. Appl. 391 535
Google Scholar
[2] Uthayakumar A, Han Y G, Lee S B 2006 Chaos Solitons Fractals 29 916
Google Scholar
[3] Bender C M, Boettcher S 1998 Phys. Rev. Lett. 80 5243
Google Scholar
[4] El-Ganainy R, Makris K G, Christodoulides D N, Musslimani Z H 2007 Opt. Lett. 32 2632
Google Scholar
[5] Bender C M 2007 Rep. Prog. Phys. 70 947
Google Scholar
[6] Baudouin L, Puel J P 2002 Inverse Probl. 18 1537
Google Scholar
[7] Avdonin S A, Mikhaylov A S, Mikhaylov V S, Park J C 2021 Inverse Problems 37 035002
Google Scholar
[8] Raissi M, Perdikaris P, Karniadakis G E 2019 J. Comput. Phys. 378 686
Google Scholar
[9] Zhou Z, Yan Z 2021 Phys. Lett. A 387 127010
Google Scholar
[10] Li J, Li B 2021 Commun. Theor. Phys. 73 125001
Google Scholar
[11] 张坤 2024 理论数学 14 117
Google Scholar
Zhang K 2024 Pure Math. 14 117
Google Scholar
[12] Qiu W X, Geng K L, Zhu B W, Liu W, Li J T, Dai C Q 2024 Nonlinear Dyn. 112 10215
Google Scholar
[13] Song J, Yan Z 2023 Physica D 448 133729
Google Scholar
[14] Wang S, Wang H, Perdikaris P 2021 Comput. Methods Appl. Mech. Eng. 384 113938
Google Scholar
[15] Liu Y, Wu R, Jiang Y 2024 J. Comput. Phys. 518 113341
Google Scholar
[16] Karniadakis G E, Kevrekidis I G, Lu L, Perdikaris P, Wang S, Yang L 2021 Nat. Rev. Phys. 3 422
Google Scholar
[17] Lu L, Meng X, Mao Z, Karniadakis G E 2021 SIAM Rev. 63 208
Google Scholar
[18] 孙志忠 2022 偏微分方程数值解法 (第三版) (北京: 科学出版社) 第 320—349页
Sun Z Z 2022 Numerical Solution of Partial Differential Equations. (3rd Ed.) (Beijing: Science Press) pp320–349
[19] Dembo R S, Eisenstat S C, Steihaug T 1982 SIAM J. Numer. Anal. 19 400
Google Scholar
[20] Göksel İ, Antar N, Bakırtaş İ 2015 Opt. Commun. 354 277
Google Scholar
-
图 1 $ V_{{\rm{PT}}}(x) $实部及虚部反演结果 (a) $ V_{{\rm{PT}}}(x) $实部的实际精确值与INTA和mPINNs反演结果的对比; (b) $ V_{{\rm{PT}}}(x) $虚部的实际精确值与INTA和mPINNs反演结果的对比; (c) INTA反演结果实部与虚部的绝对误差
Figure 1. Real and imaginary part inversion results of $ V_{{\rm{PT}}}(x) $: (a) Comparison of the real part of $ V_{{\rm{PT}}}(x) $ with the inversion results of INTA and mPINNs; (b) comparison of the imaginary part of $ V_{{\rm{PT}}}(x) $ with the inversion results of INTA and mPINNs; (c) absolute error of the real and imaginary parts of INTA inversion results.
图 2 (a) INTA每次迭代后的误差; (b) mPINNs与精确解$ u(x, z) $之间的误差
Figure 2. (a) Error after each iteration of INTA; (b) error between mPINNs and the exact solution $ u(x, z) $.
图 3 $ V_{{\rm{PT}}}(x) $实部及虚部反演结果(${\boldsymbol{ U}^*} $包含噪声) (a) INTA和mPINNs反演结果与$ V_{{\rm{PT}}}(x) $实部的实际精确值的对比; (b) INTA和mPINNs反演结果与$ V_{{\rm{PT}}}(x) $虚部的实际精确值的对比; (c) INTA反演结果实部与虚部的绝对误差
Figure 3. Real and imaginary part inversion results of $ V_{{\rm{PT}}}(x) $ ($ {\boldsymbol{U}^*} $ with noise): (a) Comparison of the real part of $ V_{{\rm{PT}}}(x) $ with the inversion results of INTA and mPINNs; (b) Comparison of the imaginary part of $ V_{{\rm{PT}}}(x) $ with the inversion results of INTA and mPINNs; (c) absolute error of the real and imaginary parts of INTA inversion results.
图 4 误差($ {\boldsymbol{U}}^* $包含噪声) (a) INTA每次迭代后的误差error; (b) mPINNs与精确解$ u(x, z) $之间的误差
Figure 4. Errors ($ {\boldsymbol{U}}^* $ with noise): (a) Error after each iteration of INTA; (b) error between mPINNs and the exact solution $ u(x, z) $.
表 1 INTA与mPINNs对比
Table 1. Comparison between INTA and mPINNs.
算法 实部误差 虚部误差 运行时间/s INTA $ 6.1768\times10^{-13} $ $ 7.6891\times10^{-12} $ 61.3 mPINNs 0.0416 0.01549 113.8 
DownLoad: CSV
表 2 绝对误差
Table 2. Absolute error.
x $ 0.1 $ $ 0.2 $ $ 0.3 $ $ 0.4 $ $ 0.5 $ $ 0.6 $ $ 0.7 $ $ 0.8 $ $ 0.9 $ INTA实部 $ 6.4\times 10^{-13} $ $ 6.5\times 10^{-13} $ $ 6.1\times 10^{-13} $ $ 6.0\times 10^{-13} $ $ 6.1\times 10^{-13} $ $ 6.4\times 10^{-13} $ $ 7.3\times 10^{-13} $ $ 7.8\times 10^{-13} $ $ 8.0\times 10^{-13} $ mPINNs实部 $ 5.3\times 10^{-3} $ $ 8.8\times 10^{-3} $ $ 8.6\times 10^{-3} $ $ 4.8\times 10^{-3} $ $ 1.5\times 10^{-3} $ $ 8.5\times 10^{-3} $ $ 1.4\times 10^{-2} $ $ 1.6\times 10^{-2} $ $ 1.1\times 10^{-2} $ INTA虚部 $ 1.4\times 10^{-12} $ $ 1.6\times 10^{-12} $ $ 1.3\times 10^{-12} $ $ 7.5\times 10^{-13} $ $ 5.2\times 10^{-13} $ $ 4.7\times 10^{-13} $ $ 6.8\times 10^{-13} $ $ 9.5\times 10^{-13} $ $ 5.5\times 10^{-13} $ mPINNs虚部 $ 1.4\times 10^{-2} $ $ 1.7\times 10^{-3} $ $ 4.0\times 10^{-3} $ $ 2.3\times 10^{-3} $ $ 4.0\times 10^{-3} $ $ 5.3\times 10^{-3} $ $ 6.1\times 10^{-3} $ $ 5.7\times 10^{-3} $ $ 4.3\times 10^{-3} $
DownLoad: CSV
表 3 INTA与mPINNs对比($ {\boldsymbol{U}^*} $包含噪声)
Table 3. Comparison between INTA and mPINNs ($ {\boldsymbol{U}^*} $ with noise).
算法 实部误差 虚部误差 运行时间/s INTA 0.0043 0.0126 71.5 mPINNs 0.0098 0.0452 112.3
DownLoad: CSV
表 4 绝对误差($ {\boldsymbol{U}}^* $包含噪声)
Table 4. Absolute error (${\boldsymbol{U}}^* $ with noise).
x 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 INTA实部 $ 1.9 \times 10^{-4} $ $ 2.3 \times 10^{-3} $ $ 4.23 \times 10^{-3} $ $ 4.43 \times 10^{-3} $ $ 2.83 \times 10^{-3} $ $ 2.8 \times 10^{-3} $ $ 3.13 \times 10^{-3} $ $ 5.23 \times 10^{-3} $ $ 3.23 \times 10^{-3} $ mPINNs实部 $ 8.93 \times 10^{-3} $ $ 1.1 \times 10^{-2} $ $ 1.2 \times 10^{-2} $ $ 1.1 \times 10^{-2} $ $ 8.53 \times 10^{-3} $ $ 5.93 \times 10^{-3} $ $ 3.53 \times 10^{-3} $ $ 1.93 \times 10^{-3} $ $ 1.33 \times 10^{-3} $ INTA虚部 $ 3.53 \times 10^{-3} $ $ 2.33 \times 10^{-3} $ $ 2.63 \times 10^{-3} $ $ 8.7 \times 10^{-4} $ $ 2.3 \times 10^{-4} $ $ 1.83 \times 10^{-3} $ $ 2.43 \times 10^{-3} $ $ 3.83 \times 10^{-3} $ $ 1.83 \times 10^{-3} $ mPINNs虚部 $ 2.43 \times 10^{-3} $ $ 6.23 \times 10^{-3} $ $ 1.0 \times 10^{-2} $ $ 1.6 \times 10^{-2} $ $ 1.6 \times 10^{-2} $ $ 1.6 \times 10^{-2} $ $ 1.5 \times 10^{-2} $ $ 1.23 \times 10^{-2} $ $ 7.13 \times 10^{-3} $
DownLoad: CSV
-
[1] Liu Y P, Gao Y T, Wei G M 2012 Phys. A Stat. Mech. Appl. 391 535
Google Scholar
[2] Uthayakumar A, Han Y G, Lee S B 2006 Chaos Solitons Fractals 29 916
Google Scholar
[3] Bender C M, Boettcher S 1998 Phys. Rev. Lett. 80 5243
Google Scholar
[4] El-Ganainy R, Makris K G, Christodoulides D N, Musslimani Z H 2007 Opt. Lett. 32 2632
Google Scholar
[5] Bender C M 2007 Rep. Prog. Phys. 70 947
Google Scholar
[6] Baudouin L, Puel J P 2002 Inverse Probl. 18 1537
Google Scholar
[7] Avdonin S A, Mikhaylov A S, Mikhaylov V S, Park J C 2021 Inverse Problems 37 035002
Google Scholar
[8] Raissi M, Perdikaris P, Karniadakis G E 2019 J. Comput. Phys. 378 686
Google Scholar
[9] Zhou Z, Yan Z 2021 Phys. Lett. A 387 127010
Google Scholar
[10] Li J, Li B 2021 Commun. Theor. Phys. 73 125001
Google Scholar
[11] 张坤 2024 理论数学 14 117
Google Scholar
Zhang K 2024 Pure Math. 14 117
Google Scholar
[12] Qiu W X, Geng K L, Zhu B W, Liu W, Li J T, Dai C Q 2024 Nonlinear Dyn. 112 10215
Google Scholar
[13] Song J, Yan Z 2023 Physica D 448 133729
Google Scholar
[14] Wang S, Wang H, Perdikaris P 2021 Comput. Methods Appl. Mech. Eng. 384 113938
Google Scholar
[15] Liu Y, Wu R, Jiang Y 2024 J. Comput. Phys. 518 113341
Google Scholar
[16] Karniadakis G E, Kevrekidis I G, Lu L, Perdikaris P, Wang S, Yang L 2021 Nat. Rev. Phys. 3 422
Google Scholar
[17] Lu L, Meng X, Mao Z, Karniadakis G E 2021 SIAM Rev. 63 208
Google Scholar
[18] 孙志忠 2022 偏微分方程数值解法 (第三版) (北京: 科学出版社) 第 320—349页
Sun Z Z 2022 Numerical Solution of Partial Differential Equations. (3rd Ed.) (Beijing: Science Press) pp320–349
[19] Dembo R S, Eisenstat S C, Steihaug T 1982 SIAM J. Numer. Anal. 19 400
Google Scholar
[20] Göksel İ, Antar N, Bakırtaş İ 2015 Opt. Commun. 354 277
Google Scholar
DownLoad:
Catalog
Metrics
- Abstract views: 310
- PDF Downloads: 15
- Cited By: 0
