Abstract
A new method named bilinear neural network is introduced in this paper, and the corresponding tensor formula is proposed to obtain the exact analytical solutions of nonlinear partial differential equations (PDEs). This is the first time that the neural network model is used to find the exact analytical solution, and this method covers almost all methods of constructing a function after bilinearization to solve nonlinear PDEs. Furthermore, this method is most likely a universal method to obtain the exact analytical solutions of nonlinear PDEs. Abundant arbitrary functions solutions of the reduced p-gBKP equation are obtained by using this method. Various beautiful plots of the presented solutions, which show diversity of exact solutions to PDEs, are made. By choosing appropriate values and functions, the fractal solitons waves are obtained and the self-similar characteristics of these waves are observed by reducing the observation range and magnifying local images. Via various three-dimensional plots, the evolution characteristics of these waves are exhibited.
Similar content being viewed by others
Explore related subjects
Discover the latest articles, news and stories from top researchers in related subjects.Avoid common mistakes on your manuscript.
1 Introduction
In order to avoid the complex calculations caused by the derivation of multilayer composite functions, neural network model has always been used to obtain numerical approximate solutions by gradient descent algorithm since its introduction. Therefore, it has not been paid attention to derive exact analytical solutions of nonlinear partial differential equations (PDEs). In recent years, with the development of various technologies and the improvement of computing power, researchers are increasingly interested in numerical and symbolic computation [1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33]. The study of nonlinear PDEs has begun since the end of the nineteenth century, and various methods have been proposed. Among them, based on bilinear transformation, the method of trying to construct a function to seeking an exact analytical solution through symbolic calculation is very hot in the nonlinear field: for example, the method to obtain the lump solution and the arbitrary function interaction solution proposed by Ma et al. [34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50]; the method to obtain the breather-type kink soliton solutions used by Sun et al. [29]; the method to obtain the periodic lump-type solutions used by Dong et al. [30]; the method to obtain the rational solutions used by Jia et al. [32]; the method to obtain the exact solutions used by Lü et al. [51,52,53,54,55,56,57,58]; the method to obtain general lump-type solutions introduced by Yong et al. [31]; the method to obtain the periodic solitary wave solutions and the three-wave solutions introduced by Liu et al. [15,16,17]; and the method to obtain the interaction solution used by Zhang et al. [22] and so on. But due to the complexity of the nonlinear equations themselves, there is no universal method to obtain the exact analytical solution for nonlinear PDEs. Because the neural network model has very complex nonlinear properties, the bilinear neural network method introduced in this paper covers all the methods listed above and will most likely be a universal method for seeking the exact analytical solution of nonlinear PDEs that may be proved based on the universal approximation theorem [59].
In this paper, bilinear neural network method would be introduced with the example of the following dimensionally reduced p-gBKP equation [34]:
where \(u_{y} = v_{x}.\) The lump solutions of Eq. (1) are obtained by Ma et al. [34], and the interaction phenomenon is discussed by Zhang et al. [8].
The rest of this paper is as follows. In Sect. 2, bilinear neural network method is introduced and the corresponding tensor formula is proposed to obtain the exact analytical solutions of nonlinear PDEs. In Sect. 3, the exact analytic solutions of Eq. (1) is obtained by using this method with the “3-2-2-1” model. The fractal soliton waves are found, and the dynamical characteristics of these waves are exhibited via various three-dimensional and density plots. That will be widely used to describe many interesting nonlinear phenomena in the fields of gas, plasma, optics, acoustics, heat transfer, fluid dynamics, classical mechanics and so on. A few of conclusions and outlook will be given in Sect. 4.
2 Bilinear neural network method and its corresponding tensor formula
2.1 Bilinear form
Under the dependent variable transformation:
Eq. (1) is transformed into the following generalized bilinear form with \(p=3\),
where p is an arbitrarily given natural number, often a prime number, and the generalized bilinear operators are defined by [39]
where \( n_{1},\ldots ,n_{M}\) are arbitrary nonnegative integers, and for an integer m, the m th power of \(\alpha \) is computed as follows:
taking \(p=3\), we have
which leads to
but if \(p=2\), then we can get,
We have noticed that when \(p = 2\), it is the Hirota bilinear operator. Transform (2) is also a characteristic transformation for establishing Bell polynomial theory of soliton equation [41]. The relation between the generalized bilinear p-gBKP Eq. (3) and the reduced p-gBKP Eq. (1) is given as follows:
Hence, if f solves the generalized bilinear p-gBKP Eq. (3), the reduced p-gBKP Eq. (1) will be solved. Similar solution procedures have been used in generating lump solutions [60,61,62] and their interactions solutions [63, 64] and even for linear PDEs in (3+1)-dimensions [65, 66].
2.2 Neural network model and corresponding tensor formula
To search for the exact analytical solutions of the bilinear p-gBKP Eq. (3), the tensor formula of nonlinear neural network is constructed as follows:
where \(W_{a,b}\) is the weight coefficient of neuron a to b. F is a generalized activation function, which can be defined arbitrarily, but in the last layer, function F must be satisfied that \(F_{{\mathbf {l_n} }} (\xi ) \ge 0\). \({\mathbf {l_n}} = \{m_{n-1}+1,m_{n-1}+2,\ldots ,n\}\) represents the nth layer space of the neural network model. \(\xi _{{\mathbf {l_i} }}\) is given as follows:
where \({\mathbf {l_0} }=\{x,y,\ldots ,t\}\), \({\mathbf {l_1} }=\{1,2,\ldots ,m_1\}\), \({\mathbf {l_i} }=\{m_{i-1}+1,m_{i-1}+2,\ldots ,m_i\}\), (\(i=2,3,\ldots ,n-1\)), b means a threshold, which can be simply understood here as an constant. This neural network tensor model can be intuitively understood through Fig. 1.
Substituting Eq. (10) into the bilinear nonlinear PDEs, a complicated equation can be obtained. Then, making the coefficient of each term in this equation equal to zero, we can obtain the algebraic equations. Solving these sets of algebraic equations by symbolic computation with the help of Maple, the coefficient solutions can be obtained. Finally, substituting these solutions and nonlinear neural network tensor formula Eq. (10) into bilinear transformation Eq. (2), the exact analytical solutions of nonlinear PDEs can be derived.
3 Abundant exact analytical solutions and the fractal soliton waves of the p-gBKP equation
To search for the exact analytical solutions of the bilinear p-gBKP Equation (3), we can choose a “3-2-2-1” neural network model, which means that there are three neurons in the input layer \({\mathbf {l_0} }\), two neurons in hidden layer \({\mathbf {l_1} }\), two neurons in hidden layer \({\mathbf {l_2} }\) and one neuron in the print layer f. This “3-2-2-1” model can be intuitively understood through Fig. 4a. By considering \({\mathbf {l_0}} = \{x,y,t\}\), \({\mathbf {l_1}} = \{1,2\}\) and \({\mathbf {l_2}} = \{3,4\}\), we procure:
where \(W_{i,j}\) (\(i=x,y,t,1,2,3,4\), \(j=1,2,3,4,f\) and \(i\ne j\)) and \(b_{k}\) (\(k=1,2,3,4\)) are real parameters to be determined later.
Substituting (12) into Eq. (3), we obtained a complicated equation. Making the coefficient of each term in this equation equal to zero, we have obtained an overdetermined nonlinear algebraic equation system with a total of 77 equations. Solving these algebraic equations by the symbolic computation with the help of Maple, we get the following four classes of solutions:
Substituting the solution of Case 1 into (12), we can get the exact analytical solution for p-gBKP equation through the bilinear transformation Eq. (2),
where the functions \(\xi _2\), \(\xi _3\) and \(\xi _4\) are given as follows:
In order to analyze the dynamics properties and discuss the evolution characteristic briefly, we could choose appropriate values and function of these parameters in Eq. (17) as
evolution characteristics of the solutions derived via the appropriate values and function list above are exhibited in Fig. 2.
Moreover, we can also choose appropriate values and function of these parameters in (17) as
dynamical characteristics of the solutions derived via the appropriate values and function list above are exhibited in Fig. 3.
Via various three-dimensional and contour plots, we find the fractal soliton waves and the self-similar characteristics of these waves are observed by reducing the observation range and magnifying local image; dynamical characteristics of these waves are exhibited in Fig. 3. That will be widely used to describe many interesting nonlinear phenomena in the fields of gas, plasma, optics, acoustics, heat transfer, fluid dynamics, classical mechanics and so on.
4 Conclusion
In this paper, a new method named bilinear neural network is introduced and the corresponding tensor formula is proposed to obtain the exact analytical solutions of nonlinear PDEs. This is the first time, the neural network model is used to solve the exact analytical solution of nonlinear PDEs.
The evolution phenomenon of the waves for p-gBKP equation is observed in Fig. 2 by choosing the appropriate parameters and functions. The fractal soliton waves are found and the self-similar characteristics of these waves are observed by reducing the observation range and magnifying local images in Fig. 3. The dynamical characteristics of these waves are exhibited via various three-dimensional and contour plots. That will be widely used to describe many interesting nonlinear phenomena in the fields of gas, plasma, optics, acoustics, heat transfer, fluid dynamics, classical mechanics and so on.
In addition, this method covers almost all methods of constructing a function after bilinearization to solve nonlinear PDEs that can be seen as models with only one hidden layer (see Fig. 4b): for example, the lump solution (which can be seen as a single hidden layer neural network model with \(F_1(\xi )=F_2(\xi )=\xi ^2, F_3(\xi )=C\)) and the arbitrary function interaction solution (which can be seen as a single hidden layer neural network model with \(F_1(\xi )=F_2(\xi )=\xi ^2\), \(F_3(\xi )\) is the arbitrary function) proposed by Ma et al. [34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50]; the breather-type kink soliton solutions method (which can be seen as a single hidden layer neural network model with \(F_1(\xi )=e^{-p_1\xi }, F_2(\xi )=\cos (p\xi ), F_3(\xi )=e^{p_1\xi }\)) used by Sun et al. [29]; the periodic lump-type solutions method (which can be seen as a single hidden layer neural network model with \(F_1(\xi )=\cosh (-p_1\xi ), F_2{\xi }=\cos (p\xi ), F_3(\xi )=\cosh (-p_1\xi )\)) used by Dong et al. [30]; the rational solutions method (which can be seen as a single hidden layer neural network model with \(F_3(\xi )=F_2(\xi )=\xi ^2, F_1(\xi )=1\)) used by Jia et al. [32]; the exact solutions method (which can be seen as a single hidden layer neural network model with \(F_1(\xi )=e^{-\xi }, F_2(\xi )=\sin (\xi )\,\,\mathrm{or}\,\,\cos (\xi ), F_3(\xi )=\sinh (\xi )\,\,\mathrm{or}\,\,\cosh (\xi ), F_4(\xi )=e^\xi \)) used by Lü et al. [57]; general lump-type solutions method (which can be seen as a single hidden layer neural network model with \(F_1=C, F_i(\xi )=\xi ^2,(i=2,3,\ldots ,n)\)) introduced by Yong et al. [31]; the periodic solitary wave solutions method (which can be seen as a single hidden layer neural network model with \(F_1(\xi )=e^{-\xi }, F_2(\xi )=\tan (\xi ), F_3(\xi )=\tanh (\xi ), F_4(\xi )=e^\xi \)) and the three-wave solutions method (which can be seen as a single hidden layer neural network model with \(F_1(\xi )=e^{-\xi }, F_2(\xi )=\cos (\xi ), F_3(\xi )=\sin (\xi ), F_4(\xi )=e^\xi \)) introduced by Liu et al. [15,16,17]; and the interaction solution method (which can be seen as a single hidden layer neural network model with \(F_1(\xi )=F_2(\xi )=\xi ^2, F_3(\xi )=e^\xi \) or \( \cosh (\xi )\)) used by Zhang et al. [22] and so on. Furthermore, the bilinear neural network method will most likely be a universal method for seeking the exact analytical solution of nonlinear PDEs that may be proved based on the universal approximation theorem [59].
References
Liu, X.Y., Triki, H., Zhou, Q., Liu, W.J., Biswas, A.: Analytic study on interactions between periodic solitons with controllable parameters. Nonlinear Dyn. 94, 703–709 (2018)
Zhang, Y., Liu, Y.P., Tang, X.Y.: M-lump and interactive solutions to a (3+1)-dimensional nonlinear system. Nonlinear Dyn. 93, 2533–2541 (2018)
Zhang, Y.J., Yang, C.Y., Yu, W.T., Mirzazadeh, M., Zhou, Q., Liu, W.J.: Interactions of vector anti-dark solitons for the coupled nonlinear Schrödinger equation in inhomogeneous fibers. Nonlinear Dyn. 94, 1351–1360 (2018)
Ankiewicz, A., Akhmediev, N.: Rogue wave-type solutions of the mKdV equation and their relation to known NLSE rogue wave solutions. Nonlinear Dyn. 91, 1931–1938 (2018)
Carboni, B., Lacarbonara, W.: Nonlinear dynamic characterization of a new hysteretic device: experiments and computations. Nonlinear Dyn. 83, 23–39 (2016)
Tang, Y.N., Tao, S.Q., Guan, Q.: Lump solitons and the interaction phenomena of them for two classes of nonlinear evolution equations. Nonlinear Dyn. 89, 429–442 (2017)
Arena, A., Lacarbonara, W.: Nonlinear parametric modeling of suspension bridges under aeroelastic forces: torsional divergence and flutter. Nonlinear Dyn. 70, 2487–2510 (2012)
Zhang, R.F., Bilige, S.D., Bai, Y.X., Lü, J.Q., Gao, X.Q.: Interaction phenomenon to dimensionally reduced p-gBKP equation. Mod. Phys. Lett. B. 32(6), 1850074 (2018)
Lü, J.Q., Bilige, S.D., Chaolu, T.: The study of lump solution and interaction phenomenon to (2+1)-dimensional generalized fifth-order kdv equation. Nonlinear Dyn. 91(2), 1669–1676 (2018)
Lü, J.Q., Bilige, S.D., Gao, X.Q., Bai, Y.X., Zhang, R.F.: Abundant lump solutions and interaction phenomena to the Kadomtsev–Petviashvili–Benjamin–Bona–Mahony equation. J. Appl. Math. Phys. 6, 1733–1747 (2018)
Lv, J.Q., Bilige, S.D.: Lump solutions of a (2+1)-dimensional bSK equation. Nonlinear Dyn. 90, 2119–2124 (2017)
Lü, J.Q., Bilige, S.D.: Diversity of interaction solutions to the (3+1)-dimensional Kadomtsev–Petviashvili–Boussinesq-like equation. Mod. Phys. Lett. B. 13, 1850311 (2018)
Wang, X.M., Bilige, S.D., Bai, Y.X.: A general sub-equation method to the burgers-like equation. Therm. Sci. 21(4), 1681–1687 (2017)
Lü, J.Q., Bilige, S.D.: The study of lump solution and interaction phenomenon to (2+1)-dimensional potential Kadomstev-Petviashvili equation. Math. Phys. Anal. (2018). https://doi.org/10.1007/s13324-018-0256-2
Liu, J.G., Du, J.Q., Zeng, Z.F., Nie, B.: New three-wave solutions for the (3+1)-dimensional Boiti–Leon–Manna–Pempinelli equation. Nonlinear Dyn. 88(1), 655–661 (2017)
Liu, J.G.: Lump-type solutions and interaction solutions for the (2+1)-dimensional generalized fifth-order KdV equation. Appl. Math. Lett. 86, 36–41 (2018)
Li, Y., Liu, J.G.: New periodic solitary wave solutions for the new (2+1)-dimensional Korteweg–de Vries equation. Nonlinear Dyn. 91(1), 497–504 (2018)
Lü, Z.S., Chen, Y.N.: Construction of rogue wave and lump solutions for nonlinear evolution equations. Eur. Phys. J. B 88(7), 88–187 (2015)
Lü, Z.S., Chen, Y.N.: Constructing rogue wave prototypes of nonlinear evolution equations via an extended tanh method. Chaos Solitons Fractals 81, 218–223 (2015)
Zhao, Z.L., Chen, Y., Han, B.: Lump soliton, mixed lump stripe and periodic lump solutions of a (2+1)-dimensional asymmetrical Nizhnik Novikov Veselov equation. Mod. Phys. Lett. B 31(14), 1750157 (2017)
Zhang, X.E., Chen, Y.: Deformation rogue wave to the (2+1)-dimensional KdV equation. Nonlinear Dyn. 90(2), 755–763 (2017)
Zhang, X.E., Chen, Y.: Rogue wave and a pair of resonance stripe solitons to a reduced (3+1)-dimensional Jimbo–Miwa equation. Commun. Nonlinear Sci. Numer. Simul. 52, 24–31 (2017)
Xu, T., Chen, Y.: Mixed interactions of localized waves in the three-component coupled derivative nonlinear Schr\(\ddot{o}\)dinger equations. Nonlinear Dyn. 92, 2133–2142 (2018)
Yang, B., Chen, Y.: Dynamics of high-order solitons in the nonlocal nonlinear Schrödinger equations. Nonlinear Dyn. 94, 489–502 (2018)
Zhang, X.E., Chen, Y.: General high-order rogue waves to nonlinear Schrödinger–Boussinesq equation with the dynamical analysis. Nonlinear Dyn. 93, 2169–2184 (2018)
Wazwaz, A.M.: Two-mode fifth-order kdv equations: necessary conditions for multiple-soliton solutions to exist. Nonlinear Dyn. 87(3), 1685–1691 (2017)
Osman, M.S., Wazwaz, A.M.: An efficient algorithm to construct multi-soliton rational solutions of the (2+1)-dimensional kdv equation with variable coefficients. Appl. Math. Comput. 321, 282–289 (2018)
Wazwaz, A.M.: Compact and noncompact physical structures for the ZK–BBM equation. Appl. Math. Comput. 169(1), 713–725 (2017)
Sun, Y., Tian, B., Xie, X.Y., Chai, J., Yin, H.H.: Rogue waves and lump solitons for a-dimensional B-type Kadomtsev–Petviashvili equation in fluid dynamics. Waves Random Complex Media 28(3), 544–552 (2018)
Dong, M.J., Tian, S.F., Wang, X.B., Zhang, T.T.: Lump-type solutions and interaction solutions in the (3+1)-dimensional potential Yu–Toda–Sasa–Fukuyama equation. Anal. Math. Phys. (2018). https://doi.org/10.1007/s13324-018-0258-0
Yong, X.L., Li, X.J., Huang, Y.H.: General lump-type solutions of the (3+1)-dimensional Jimbo–Miwa equation. Appl. Math. Lett. 86, 222–228 (2018)
Jia, S.L., Gao, Y.T., Hu, L., Huang, Q.M., Hu, W.Q.: Soliton-like periodic wave and rational solutions for a (3+1)-dimensional Boiti–Leon–Manna–Pempinelli equation in the incompressible fluid. Superlattices Microstruct. 102, 273–283 (2017)
Zhang, Y., Dong, H.H., Zhang, X.E., Yang, H.W.: Rational solutions and lump solutions to the generalized (3+1)-dimensional shallow water-like equation. Comput. Math. Appl. 73, 246–252 (2017)
Ma, W.X., Qin, Z.Y., Lü, X.: Lump solutions to dimensionally reduced p-gKP and p-gBKP equations. Nonlinear Dyn. 84, 923–931 (2016)
Ma, W.X., Zhou, Y.: Lump solutions to nonlinear partial differential equations via Hirota bilinear forms. J. Differ. Equ. 264(4), 2633–2659 (2018)
Zhang, J.B., Ma, W.X.: Mixed lump-kink solutions to the BKP equation. Comput. Math. Appl. 74, 591–596 (2017)
Yang, J.Y., Ma, W.X., Qin, Z.Y.: Lump and lump-soliton solutions to the (2+1)-dimensional Ito equation. Anal. Math. Phys. 8, 427–436 (2018)
Ma, W.X., Yong, X., Zhang, H.Q.: Diversity of interaction solutions to the (2+1)-dimensional Ito equation. Comput. Math. Appl. 75, 289–295 (2018)
Ma, W.X.: Generalized bilinear differential equations. Stud. Nonlinear Sci. 2(4), 140–144 (2011)
Ma, W.X.: Lump solutions to the Kadomtsev–Petviashvili equation. Phys. Lett. A 379(36), 1975–1978 (2015)
Ma, W.X.: Bilinear equations, Bell polynomials and linear superposition principal. J. Phys. 411, 12021 (2013)
Ma, W.X., Yong, X., Zhang, H.Q.: Diversity of interaction solutions to the (2+1)-dimensional ito equation. Comput. Math. Appl. 75(1), 289–295 (2018)
Zhang, H.Q., Ma, W.X.: Lump solutions to the (2+1)-dimensional Sawada–Kotera equation. Nonlinear Dyn. 87(4), 2305–2310 (2017)
Ma, W.X.: Lump-type solutions to the (3+1)-dimensional Jimbo–Miwa equation. Int. J. Nonlinear Sci. Numer. 17, 355–359 (2016)
Lü, X., Ma, W.X., Khalique, C.M.: A direct bilinear Bäcklund transformation of a (2+1)-dimensional Korteweg–de Vries-like model. Appl. Math. Lett. 50, 37–42 (2015)
Lü, X., Ma, W.X., Yu, J., Khalique, C.M.: Solitary waves with the Madelung fluid description: a generalized derivative nonlinear Schrödinger equation. Commun. Nonlinear Sci. Numer. Simul. 31, 40–46 (2016)
Lü, X., Ma, W.X., Yu, J., Lin, Fh, Khalique, C.M.: Envelope bright-soliton and dark-soliton solutions for the Gerdjikov–Ivanov model. Nonlinear Dyn. 82, 1211–1220 (2015)
Lü, X., Ma, W.X., Zhou, Y., Khalique, C.M.: Rational solutions to an extended Kadomtsev–Petviashvili-like equation with symbolic computation. Comput. Math. Appl. 71, 1560–1567 (2016)
Lü, X., Ma, W.X.: Study of lump dynamics based on a dimensionally reduced Hirota bilinear equation. Nonlinear Dyn. 85, 1217–1222 (2016)
Lü, X., Ma, W.X., Chen, S.T., Chaudry, M.K.: A note on rational solutions to a Hirota–Satsuma-like equation. Appl. Math. Lett. 58, 13–18 (2016)
Lü, X., Chen, S.T., Ma, W.X.: Constructing lump solutions to a generalized Kadomtsev–Petviashvili–Boussinesq equation. Nonlinear Dyn. 86, 523–534 (2016)
Gao, L.N., Zhao, X.Y., Zi, Y.Y., Lü, X.: Resonant behavior of multiple wave solutions to a Hirota bilinear equation. Comput. Math. Appl. 72, 1225–1229 (2016)
Gao, L.N., Zi, Y.Y., Yin, Y.H., Ma, W.X., Lü, x: Bäcklund transformation, multiple wave solutions and lump solutions to a (3+1)-dimensional nonlinear evolution equation. Nonlinear Dyn. 89, 2233–2240 (2017)
Lü, X., Lin, F.H.: Soliton excitations and shape-changing collisions in alphahelical proteins with interspine coupling at higher order. Commun. Nonlinear Sci. 32, 241–261 (2016)
Lin, F.H., Chen, S.T., Qu, Q.X., Wang, J.P., Zhou, X.W., Lü, X.: Resonant multiple wave solutions to a new (3+1)-dimensional generalized Kadomtsev–Petviashvili equation: linear superposition principle. Appl. Math. Lett. 78, 112–117 (2018)
Lü, X., Wang, J.P., Lin, F.H., Zhou, X.W.: Lump dynamics of a generalized two-dimensional Boussinesq equation in shallow water. Nonlinear Dyn. 91(2), 1249–1259 (2018)
Yin, Y.H., Ma, W.X., Liu, J.G., Lü, X.: Diversity of exact solutions to a (3+1)-dimensional nonlinear evolution equation and its reduction. Comput. Math. Appl. 76, 1275–1283 (2018)
Batwa, S., Ma, W.X.: A study of lump-type and interaction solutions to a (3+1)-dimensional Jimbo–Miwa-like equation. Comput. Math. Appl. 76, 1576–1582 (2018)
Hornik, K.: Approximation capabilities of multilayer feedforward networks. Neural Netw. 4, 251–257 (1991)
Chen, S.T., Ma, W.X.: Lump solutions to a generalized Bogoyavlensky–Konopelchenko equation. Front. Math. China 13(3), 525–534 (2018)
Chen, S.T., Ma, W.X.: Lump solutions of a generalized Calogero–Bogoyavlenskii–Schiff equation. Comput. Math. Appl. 76, 1680–1685 (2018)
Manukure, S., Zhou, Y., Ma, W.X.: Lump solutions to a (2+1)-dimensional extended KP equation. Comput. Math. Appl. 75, 2414–2419 (2018)
Zhao, H.Q., Ma, W.X.: Mixed lump-kink solutions to the KP equation. Comput. Math. Appl. 74, 1399–1405 (2017)
Yang, J.Y., Ma, W.X., Qin, Z.Y.: Abundant mixed lump-soliton solutions to the BKP equation. East Asian J. Appl. Math. 8(2), 224–232 (2018)
Ma, W.X.: Lump and interaction solutions of linear PDEs in (3+1)-dimensions. East Asian J. Appl. Math. https://doi.org/10.4208/eajam.100218.300318
Ma, W.X.: Abundant lumps and their interaction solutions of (3+1)-dimensional linear PDEs. J. Geom. Phys. 133, 10–16 (2018)
Acknowledgements
This work is supported by the National Natural Science Foundation of China (11661060) and the Natural Science Foundation of Inner Mongolia Autonomous Region of China (2018LH01013).
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
Conflict of interest
The authors declare that they have no conflict of interest concerning the publication of this manuscript.
Rights and permissions
About this article
Cite this article
Zhang, RF., Bilige, S. Bilinear neural network method to obtain the exact analytical solutions of nonlinear partial differential equations and its application to p-gBKP equation. Nonlinear Dyn 95, 3041–3048 (2019). https://doi.org/10.1007/s11071-018-04739-z
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11071-018-04739-z