Peanut yield prediction with UAV multispectral imagery using a cooperative machine learning approach

Tej Bahadur Shahi; Cheng-Yuan Xu; Arjun Neupane; Dayle B. Fleischfresser; Daniel J. O'Connor; Graeme C. Wright; William Guo; Tej Bahadur Shahi; Cheng-Yuan Xu; Arjun Neupane; Dayle B. Fleischfresser; Daniel J. O'Connor; Graeme C. Wright; William Guo

doi:10.3934/era.2023169

Electronic Research Archive

2023, Volume 31, Issue 6: 3343-3361. doi: 10.3934/era.2023169

Previous Article Next Article

Research article Special Issues

Peanut yield prediction with UAV multispectral imagery using a cooperative machine learning approach

1.
School of Engineering and Technology, Central Queensland University, Rockhampton, Australia
2.
School of Health, Medical, and Applied Sciences, Central Queensland University, Bundaberg, Australia
3.
Queensland Department of Agriculture and Fisheries, Townsville, Australia
4.
Peanut Company of Australia, Kingaroy, Australia

Received: 07 February 2023 Revised: 19 March 2023 Accepted: 30 March 2023 Published: 14 April 2023

The unmanned aerial vehicle (UAV), as a remote sensing platform, has attracted many researchers in precision agriculture because of its operational flexibility and capability of producing high spatial and temporal resolution images of agricultural fields. This study proposed machine learning (ML) models and their ensembles for peanut yield prediction using UAV multispectral data. We utilized five bands (red, green, blue, near-infra-red (NIR) and red-edge) multispectral images acquired at various growth stages of peanuts using UAV. The correlation between spectral bands and yield was analyzed for each growth stage, which showed that the maturity stages had a significant correlation between peanut yield and spectral bands: red, green, NIR and red edge (REDE). Using these four bands spectral data, we assessed the potential for peanut yield prediction using multiple linear regression and seven non-linear ML models whose hyperparameters were optimized using simulated annealing (SA). The best three ML models, random forest (RF), support vector machine (SVM) and XGBoost, were then selected to construct a cooperative yield prediction framework with both the best ML model and the ensemble scheme from the best three as comparable recommendations to the farmers.

Keywords:

Citation: Tej Bahadur Shahi, Cheng-Yuan Xu, Arjun Neupane, Dayle B. Fleischfresser, Daniel J. O'Connor, Graeme C. Wright, William Guo. Peanut yield prediction with UAV multispectral imagery using a cooperative machine learning approach[J]. Electronic Research Archive, 2023, 31(6): 3343-3361. doi: 10.3934/era.2023169

Related Papers:

[1]	Musawa Yahya Almusawa, Hassan Almusawa . Numerical analysis of the fractional nonlinear waves of fifth-order KdV and Kawahara equations under Caputo operator. AIMS Mathematics, 2024, 9(11): 31898-31925. doi: 10.3934/math.20241533
[2]	Aslı Alkan, Halil Anaç . The novel numerical solutions for time-fractional Fornberg-Whitham equation by using fractional natural transform decomposition method. AIMS Mathematics, 2024, 9(9): 25333-25359. doi: 10.3934/math.20241237
[3]	Hayman Thabet, Subhash Kendre, James Peters . Travelling wave solutions for fractional Korteweg-de Vries equations via an approximate-analytical method. AIMS Mathematics, 2019, 4(4): 1203-1222. doi: 10.3934/math.2019.4.1203
[4]	Muath Awadalla, Abdul Hamid Ganie, Dowlath Fathima, Adnan Khan, Jihan Alahmadi . A mathematical fractional model of waves on Shallow water surfaces: The Korteweg-de Vries equation. AIMS Mathematics, 2024, 9(5): 10561-10579. doi: 10.3934/math.2024516
[5]	M. Mossa Al-Sawalha, Rasool Shah, Adnan Khan, Osama Y. Ababneh, Thongchai Botmart . Fractional view analysis of Kersten-Krasil'shchik coupled KdV-mKdV systems with non-singular kernel derivatives. AIMS Mathematics, 2022, 7(10): 18334-18359. doi: 10.3934/math.20221010
[6]	Ihsan Ullah, Aman Ullah, Shabir Ahmad, Hijaz Ahmad, Taher A. Nofal . A survey of KdV-CDG equations via nonsingular fractional operators. AIMS Mathematics, 2023, 8(8): 18964-18981. doi: 10.3934/math.2023966
[7]	Muhammad Imran Liaqat, Sina Etemad, Shahram Rezapour, Choonkil Park . A novel analytical Aboodh residual power series method for solving linear and nonlinear time-fractional partial differential equations with variable coefficients. AIMS Mathematics, 2022, 7(9): 16917-16948. doi: 10.3934/math.2022929
[8]	Khalid Khan, Amir Ali, Manuel De la Sen, Muhammad Irfan . Localized modes in time-fractional modified coupled Korteweg-de Vries equation with singular and non-singular kernels. AIMS Mathematics, 2022, 7(2): 1580-1602. doi: 10.3934/math.2022092
[9]	Sayed Saifullah, Amir Ali, Zareen A. Khan . Analysis of nonlinear time-fractional Klein-Gordon equation with power law kernel. AIMS Mathematics, 2022, 7(4): 5275-5290. doi: 10.3934/math.2022293
[10]	Asif Khan, Tayyaba Akram, Arshad Khan, Shabir Ahmad, Kamsing Nonlaopon . Investigation of time fractional nonlinear KdV-Burgers equation under fractional operators with nonsingular kernels. AIMS Mathematics, 2023, 8(1): 1251-1268. doi: 10.3934/math.2023063

Abstract

1. Introduction

Currently, fractional-order calculus (FOC) plays a dynamic role in the development and applications of modern science and technology ^[1]. The FOC is a field of mathematics that deals with differentiation and integration under an arbitrary order of the operation; the order is not limited to the integer one but can also be any real or even complex number ^[2,3,4]. For a long time, it was ignored due to the integral complexity and freedom of integer order calculus, as well as the reality that FOC does not have entirely conventional physical or geometrical understandings ^[5]. Currently, there is a strong interest in generalising integer-order NPDEs to model complex engineering problems such as bio-engineering, electronics, viscoelasticity, control theory, thermal sciences, electrical networking, chemical, and fluid dynamics ^[6,7].

Korteweg-de Vries (KdV) equation has been extensively investigated in physics to describe the interaction and evaluation of non-linear waves. It was originally derived by D. J. Korteweg and G. de Vries, in an investigation to re-generate the solitary waves that govern a small amplitude in one dimension with lengthy propagating gravity waves spreading in a shallow channel of water ^[8]. Apart from solitary waves, the KdV equation has also been studied for numerous physical phenomena, including hydro-magnetic waves, lattice dynamics, physics of plasma such as electron-acoustic (EA) and ion-acoustic (IA) waves ^[9]. It is also used in aerodynamics, fluid dynamics, transportation of mass, boundary layer behavior, and modeling the mechanical behavior of materials modeled as a continuous mass for shock wave formations. The concept of excitation of the electron-acoustic (EA) mode was initially proposed by Fried and Gould in the reference ^[10]. They have noted that the Landau's dampening effect on the EA potentials weakens with increasing wave number. In further experiments, weak dampening of the EA in plasma containing both high- and low-temperature electrons was found to be caused by ^[11]. Such plasma circumstances had previously been noted in a number of contexts, including ^[12]. Iwamoto ^[13] has studied the development of the high frequency Langmuir mode and the electron-acoustic wave evolution in nonrelativistic electron-positron plasma. It has been demonstrated that, in compared to the Langmuir wave, the low frequency EA excitation Landau damped with a substantially higher growth rate. The propagation of electron acoustic waves (EAWs) in plasmas and in space environments has drawn a lot of interest due to its value in comprehending a variety of collective processes in laboratory equipment ^[14,15]. Irfan et al. ^[16] studied fast (Langmuir) and slow (electron-acoustic) modes in dense electron-positron-ions (EPI) plasma. They observed that there is a chance that the fast electron-acoustic (FEA) mode could occur. The slow electron-acoustic (SEA) solitons are not allowed to propagate because they have a negative phase dispersion impact. The EPI plasma may allow for the evolution of electron-holes at a relatively low positron concentration, leading to the formation of compressive FEA solitons. Broadband electrostatic noise (BEN) emissions in the auroral and other regions of the magnetosphere, such as the plasma sheet boundary layer (PSBL) and the polar cusp, have been detected by satellite data. The Korteweg-de Vries (KdV) equation is obtained from studies of small-amplitude EAWs in unmagnetized plasma using first-order reductive perturbation theory ^[17]. More than 20 percent of the solitary waves' amplitude is understated by the first order soliton solution. Consequently, higher-order fixes may be used to help solve this issue. Shewy et al. ^[18,19] investigated the contributions of higher-order corrections to the properties of electrostatic EAWs as well as the impacts of non-thermal distribution of hot electrons. The higher-order correction contribution is given by the time-fractional modified Korteweg-de Vries (TF-mKdV) which modulates the amplitude of the solitary wave.

The time-fractional modified KdV (TF-mKdV) equation is well-known for its contribution in the creation of the Lax pair and an unlimited quantity of conservation laws for the KdV equation ^[20]. The finite-gap integration approach and Whitham modulation theory are used to investigate the comprehensive classification of solutions to the defocusing complex modified KdV equation with step-like initial condition ^[21]. Wang et al. ^[22] employed the nonlinear steepest descent method of Deift and Zhou, the long-time asymptotics of the focusing Kundu-Eckhaus equation with nonzero boundary conditions at infinity. The TF-mKdV equation is obtained using perturbation expansions and the theory that the soliton width is modest in comparison to the plasma inhomogeneity scale length. In this case, the soliton retains its whole identity, as well as its amplitude, breadth, and speed. Further, Breathers and the unique soliton behavior of the TF-mKdV equation are well-known. Here, we study the localized solutions for time-fractional Korteweg-de Vries (TF-KdV) and time-fractional modified Korteweg-de Vries (TF-mKdV) with help of an analytical method called Aboodh transform decomposition method (ADM) given in ^[23].

From the conventional Fourier integral, the Aboodh transform is derived. The Aboodh transform was first proposed by Khalid Aboodh ^[24] to show how to solve various ordinary differential equations in the time domain. This was done because of the Aboodh transform's mathematical features and simplicity. Differential equations are typically solved using Fourier, Laplace, Elzaki, and Sumudu transforms as the primary mathematical methods. Motivated from these study we are interesting to apply the Aboodh transform together with decomposition method to find an approximate solution of our proposed model. The proposed models have been extensively used in modeling complex physical systems such as thermal as well as current flows in electric circuits. It also plays an important role in modeling particles vibrating in their lattices. It should be noted that the physical phenomena in the KdV equation will be considered as non-conservative to describe fractional differential equations. Recently, the time-fractional KdVB equation together with nonlinearities has been extensively investigated using various methods ^[25]. We consider the integer order of the time derivative term of the mKdV model given in ^[26] be in the arbitrary order of the following form

$\begin{eqnarray} &&\frac{\partial^\alpha u}{\partial t^\alpha}+a\, u^m\frac{\partial u}{\partial x}+b\, \frac{\partial^{3}u}{\partial x^3} = 0, \, \, \, \, 0 < \alpha\leq 1, \end{eqnarray}$

(1.1)

with

$\begin{eqnarray} u(x, 0) = f(x), \end{eqnarray}$

(1.2)

the numbers $a, b \in \mathbb{R}$ with $a$ is the non-linearity coefficient and $b$ is the dispersive coefficient of Eq (1.1) and the time fractional-order ( $0 < \alpha\leq 1$ ) derivative is supposed to be in Caputo's form. Equation (1.1) is known as time-fractional modified Korteweg-de Vries (TF-mKdV) equation and is used vastly in plasma physics to study solitary and shock waves ^[27]. When $m = 1$ , the TF-mKdV reduce to simple time fractional Korteweg-de Vries (TF-KdV) equation.

Recently, various descriptions for fractional-order operators are also broadly studied, for example, Caputo-Fabrizio (C-F), Liouville-Caputo, and Riemann-Liouville (R-L) ^[28]. The R-L derivatives involve the complexity of the assumed function with the power-law kernel, while, the LC derivatives include the complexity of the confined derivative of a considered problem with power-law. Recently, the fractional KdV type equations together with several other nonlinear systems involving the time-fractional order derivatives along various methods for their analytical solutions have been investigated. To investigate analytical as well as numerical solutions to FNPDEs/TF-mKdV equations, various techniques have been studied ^[29,30,31]. Here, we apply the Aboodh decomposition method (ADM) ^[23] to find an approximate solution of Eq (1.1) in Caputo's sense. The rest of the article is prepared as follows. The preliminaries section 2 includes the main definitions, remarks, and some important results regarding the proposed method. In section 3, the Aboodh transform and Aboodh decomposition method are discussed. In section 4, uniqueness and convergence analysis of the method is discussed. In section 5, we consider two examples in Caputo's form of TF-KdV and TF-mKdV equations with the application of the proposed method. This section also includes numerical solutions and discussion. In section 5, the accuracy and effectiveness of the ADM are confirmed by using the error analysis. We include the summary in section 6, while in section 7 we have given the future work and extension of our present work.

2. Preliminaries

In this section, we provide some basic definitions, theorems, lemmas, and remarks, which are essential for our proposed method. We also define some basic rules and definitions related to the Aboodh transform and decomposition methods with some related properties. The Aboodh transform (see the reference ^[32]) is defined for the functions of exponential order in the following set

$\begin{eqnarray} \mathcal{A} = \left\{u(t):\exists\, \, M, k_1, k_2, \, \, \left|u(t)\right|\le M\, e^{-vt}, \, \ k_1\leq v\leq k_2\right\}, \end{eqnarray}$

(2.1)

the constant $M$ must be a finite value for a given function $u(t)$ in the set $\mathcal{A}$ , while $k_1, k_2 \in \mathbb{R}$ . The Aboodh transform $\mathcal{A}$ $u(t)$ is given by

$\begin{eqnarray} \mathcal{A}(u(t))(v) = K(v) = \frac{1}{v}\int_{0}^{\infty}\, u(t)e^{-vt}dt, \quad\quad \, \, \, t\geq 0, \end{eqnarray}$

(2.2)

where $v \in [k_1, k_2]$ is the variable utilised to factor the variable $t$ in the function $u(.)$ .

Definition 1. ^[33] The Aboodh transform for two variables function $x\geq 0$ , $t\geq 0$ is defined by

$\begin{eqnarray} \mathcal{A}(u(x, t))(u, v) = K(u, v) = \frac{1}{uv}\int_{0}^{\infty}\int_{0}^{\infty} e^{-ut}\, e^{-vt}u(x, t)dxdt, \end{eqnarray}$

(2.3)

while the inverse Aboodh transform is give by

$\begin{equation*} u(x, t) = \mathcal{A}^{-1}\{K(u, v)\} = \frac{1}{2\pi\, i}\int\limits_{\beta-i\infty}^{\beta+i\infty}u\, e^{ut}\bigg[\frac{1}{2\pi\, i} \int\limits_{\gamma-i\infty}^{\gamma+i\infty}ve^{vx}K(u, v)\, dv\bigg]du, \end{equation*}$

where, $K(u, v)$ , is a function such that $u$ and $v$ defined such that $Re$ $(u)\geq \beta$ and $Re$ $(v)\geq \gamma$ . It should be noted that $\beta, \gamma \in \mathbb{R}$ to be considered accordingly.

Definition 2. ^[34] The Caputo's derivative for $\alpha > 0$ for $u(t)$ is defined by function $u : (0, \infty)\rightarrow {\rm I\!R}$ is given by

$\begin{eqnarray} ^cD^\alpha u(t) = \frac{1}{\Gamma{(n-\alpha)}}\int_{0}^{t}(t-s)^{n-\alpha-1}u^n(s)ds, \end{eqnarray}$

(2.4)

where $\alpha \in (n-1, n]$ , with $n = [\alpha] + 1, [\alpha]$ represent the integer part of a real number $\alpha$ , providing that the right-hand side of the above integral is continuously well-defined on $(0, \infty)$ .

Definition 3. ^[35] Applying Aboodh transform on function $u(x, t)$ in the Caputo's sense is given by

$\begin{equation} \mathcal{A}\{^cD^{\alpha}u(x, t)\} = v^\alpha K(u, v)-\sum\limits_{k = 0}^{n-1}\frac{u^{(k)}(x, 0)}{v^{2-\alpha+k}}. \end{equation}$

(2.5)

Definition 4. ^[35] The Aboodh transform of some other functions and derivatives are given as

$\begin{eqnarray*} \mathcal{A}\{t^n\} = \frac{n!}{v^{n+2}}, \quad n\in N, \, \, \, \, \mathcal{A}\{t^\alpha\} = \frac{\Gamma(\alpha+1)}{v^{\alpha+2}}, \quad \alpha \ge -1. \end{eqnarray*}$

3. Aboodh transform decomposition method

Here, we introduce ADM applied to attain a series of solutions of NODEs and NPDEs. It is a very much effective technique to find an approximate solution of dynamical systems. Writing Eq (1.1) in Caputo's sense, we obtain

$\begin{equation} \begin{aligned} ^cD_t^\alpha\, u+a\, u^m\frac{\partial u}{\partial x}+b\, \frac{\partial ^3u}{\partial x^3} = 0, \end{aligned} \end{equation}$

(3.1)

$0 < \alpha\leq 1$ and

$\begin{eqnarray} u(x, 0) = f(x). \end{eqnarray}$

(3.2)

Using ADM and the definitions given in the previous section, we obtain

$\begin{equation} \mathcal{A}\{^cD_t^\alpha\, u\}+a\mathcal{A}\{u^m\frac{\partial u}{\partial x}\}+b\mathcal{A}\{\frac{\partial^3 u}{\partial x^3}\} = 0. \end{equation}$

(3.3)

Applying the technique given in section 2 for ADM on fractional order, we obtain

$\begin{equation} \mathcal{A}\{u(x, t)\} = \frac{1}{v^2}\mathcal{A}\{u(x, 0)\}-a\frac{1}{v^\alpha}\mathcal{A} \{u^m\frac{\partial u}{\partial x}\}-b\frac{1}{v^\alpha}\mathcal{A}\{\frac{\partial^3 u}{\partial x^3}\}, \end{equation}$

(3.4)

where Eq (3.2) becomes

$\begin{eqnarray} \mathcal{A}\{u(x, 0)\} = K(u). \end{eqnarray}$

(3.5)

By considering the series solution

$\begin{equation} u = \sum\limits_{n = 0}^\infty u_n. \end{equation}$

(3.6)

The non-linearity can be obtained as

$\begin{eqnarray*} u^mu_x(x, t) = \sum\limits_{i = 0}^\infty A_n. \end{eqnarray*}$

The value of $A_n$ for the functions $u_i, \, i = 1, 2, 3, \dots$ , is given by ^[36]

$\begin{equation} A_n = \frac{1}{n!}\frac{d^n}{d\lambda^n}\left[\sum\limits_{k = 0}^n \lambda^ku^m_k(x, t)\sum\limits_{k = 0}^n \lambda^ku_{kx}(x, t)\right]_{\lambda = 0}. \end{equation}$

(3.7)

Applying an inverse Aboodh transform to above equation, we obtain

$\begin{equation} \begin{split} u_0& = \mathcal{A}^{-1}\{\frac{1}{v^2}K(u, 0)\} = u(x, 0).\\ u_1& = -a\mathcal{A}^{-1} \left\{\frac{1}{v^\alpha}\mathcal{A}\{A_0\}\right\} -b\mathcal{A}^{-1}\left\{\frac{1}{v^\alpha}\mathcal{A}\{u_{0xxx}\}\right\}, \\ u_2& = -a\mathcal{A}^{-1}\left\{\frac{1}{v^\alpha}\mathcal{A}\{A_1\}\right\}- b\mathcal{A}^{-1} \left\{\frac{1}{v^\alpha}\mathcal{A}\{u_{1xxx}\}\right\}, \\ u_3& = -a\mathcal{A}^{-1}\left\{\frac{1}{v^\alpha}\mathcal{A}\{A_2\}\right\}- b\mathcal{A}^{-1} \left\{\frac{1}{v^\alpha}\mathcal{A}\{u_{2xxx}\}\right\}, \\ &\vdots\\ u_{n+1}& = -a\mathcal{A}^{-1}\left\{\frac{1}{v^\alpha}\mathcal{A}\{A_n\}\right\}- b\mathcal{A}^{-1}\left\{\frac{1}{v^\alpha}\mathcal{A}\{u_{nxxx}\}\right\}. \end{split} \end{equation}$

(3.8)

The general solution can be written as

$\begin{equation} u(x, t) = u_0(x, t)-a\mathcal{A}^{-1}\left\{\frac{1}{v^\alpha}\mathcal{A}\{A_n\}\right\}- b\mathcal{A}^{-1}\left\{\frac{1}{v^\alpha}\mathcal{A} \{u_{nxxx}\}\right\}. \end{equation}$

(3.9)

4. Uniqueness and convergence of the method

The necessary condition that ensures the presence of a unique solution is studied in this section, and the convergence of the solutions is discussed, we follow these theorems applied to the series solutions in ^[37,38].

Theorem 1. (Uniqueness theorem) For $0 < \gamma < 1$ , the solution (3.9) of Eq (3.1) is a unique solution. where

$\begin{eqnarray} \gamma = \frac{(\ell_1+\ell_2+\ell_3)t^{\alpha+1}}{\Gamma(\alpha)}, \end{eqnarray}$

(4.1)

where $\ell_1$ and $\ell_2$ , are Lipschitz constants, $0 < \alpha\leq 1$ and $\Gamma$ is the well known Gamma function can be defined as

$\Gamma(\alpha) = \int_0^\infty e^{-t} t^{\alpha-1} dt, \quad \mathbb{R} > 0.$

Proof. Define a mapping $\mathscr{D}$ : $\mathscr{B}$ $\rightarrow$ $\mathscr{B}$ with $\mathscr{B}$ = $(C [J], \left\lVert{.}\right\rVert)$ is the Banach space on $J = [0, T]$ . We can write Eq (3.9) as

$\begin{equation*} u_{n+1} = u(x, 0)+\mathcal{A}^{-1} \left\{\frac{1}{v^\alpha}\mathcal{A}\bigg(\mathscr{H}u_n(x, t)+\mathscr{M}u_n(x, t)+\mathscr{N}u_n(x, t)\bigg)\right\}, \end{equation*}$

where $\mathscr{H}u = \frac{\partial u}{\partial x}$ , $\mathscr{M}u = \frac{\partial^3u }{\partial x^3}$ and $\mathscr{N}u = u\frac{\partial u}{\partial x}$ , and suppose that $\mathscr{H}u$ , and $\mathscr{M}u$ are also Lipschitzian with

$\begin{eqnarray*} \left|\mathscr{H}u-\mathscr{H}\bar{u}\right| < \ell_1\left|u-\bar{u}\right|, \, \, \, \, \left|\mathscr{M}u-\mathscr{M}\bar{u}\right| < \ell_2\left|u-\bar{u}\right|, \end{eqnarray*}$

where $u, \bar{u}$ are the function's distinct values. Now we proceed as follow

$\begin{array}{l} \left\lVert{\mathscr{D}u\, -\mathscr{D}\bar{u}\, }\right\rVert = \max\limits_{t\in J}\left|\mathcal{A}^{-1} \left\{\frac{1}{v^\alpha}\mathcal{A}\bigg(\mathscr{H}u_n\, +\mathscr{M}u_n\, +\mathscr{N}u_n\, \bigg)\right\}- \mathcal{A}^{-1}\\ \left\{\frac{1}{v^\alpha}\mathcal{A}\bigg(\mathscr{H}\bar{u_n}\, +\mathscr{M}\bar{u_n}\, +\mathscr{N}\bar{u_n}\, \bigg)\right\}\right|\\ \leq\max\limits_{t\in J}\Bigg|\mathcal{A}^{-1} \left\{\frac{1}{v^\alpha}\mathcal{A}\bigg(\mathscr{H}u_n\, -\mathscr{H}\bar{u}\, \bigg)\right\} +\mathcal{A}^{-1} \left\{\frac{1}{v^\alpha}\mathcal{A}\bigg(\mathscr{M}u\, -\mathscr{M}\bar{u}\, \bigg)\right\} \\ +\mathcal{A}^{-1} \left\{\frac{1}{v^\alpha}\mathcal{A}\bigg(\mathscr{N}u\, -\mathscr{N}\bar{u}\, \bigg)\right\}\Bigg| \\ \leq\max\limits_{t\in J}\Bigg|\ell_1\mathcal{A}^{-1} \left\{\frac{1}{v^\alpha}\mathcal{A}\bigg(u\, -\bar{u}\, \bigg)\right\} +\ell_2\mathcal{A}^{-1} \left\{\frac{1}{v^\alpha}\mathcal{A}\bigg(u\, -\bar{u}\, \bigg)\right\} \\ +\ell_3\mathcal{A}^{-1} \left\{\frac{1}{v^\alpha}\mathcal{A}\bigg(\mathscr{N}u\, -\bar{u}\, \bigg)\right\}\Bigg| \\ \leq\max\limits_{t\in J}(\ell_1+\ell_2+\ell_3) \Bigg|\mathcal{A}^{-1} \left\{\frac{1}{v^\alpha}\mathcal{A}\bigg(u\, -\bar{u}\, \bigg)\right\}\Bigg| \\ = \frac{(\ell_1+\ell_2+\ell_3)t^{(\alpha-1)}}{\Gamma(\alpha)} \left\lVert{u\, -\bar{u}\, }\right\rVert. \end{array}$

The mapping is contraction under the condition $0 < \gamma < 1$ . As a result of the Banach fixed point theorem for contraction, Eq (3.1) has a unique solution.

Next, we discuss convergence analysis of the problems.

Theorem 2. (Convergence theorem) The solution of Eq (3.1) in general form will be convergence.

Proof. The Banach space of all continuous functions on the interval $J$ with the norm $\left\lVert{u(x, t)}\right\rVert$ = $\max_{t\in J}\left\lVert{u(x, t)}\right\rVert$ is denoted as $(C[J], \left\lVert{.}\right\rVert)$ . Define the $\left\{\mathscr{S} n\right\}$ sequence of partial sums that is $\mathscr{S}_n$ = $\sum_{j = 0}^{n}$ . With $n\geq m$ , let $\mathscr{S}_n$ and $\mathscr{S}_m$ be arbitrary partial sums. In this Banach space, we will show that $\mathscr{S}_n$ is a Cauchy sequence. This can obtained by employing a new formulation of Adomian polynomials.

$P(\mathscr{S}_n) = \bar{A}_{n}+\sum\limits_{r1 = 0}^{n-1}\bar{A}_{{r_1}}, \quad Q(\mathscr{S}_n) = \bar{A}_{n}+\sum\limits_{r2 = 0}^{n-1}\bar{A}_{{r_2}}.$

Now

$\begin{array}{l} \left\lVert{\mathscr{S}_n-\mathscr{S}_m}\right\rVert = \left|\sum\limits_{j = 0}^nu_j\, -\sum\limits_{k = 0}^mu_k\, \right| = \max\limits_{t\in J}\left|\sum\limits_{j = m+1}^nu_j\, \right|\\ \leq\max\limits_{t\in J}\Bigg|\mathcal{A}^{-1} \left\{\frac{1}{v^\alpha}\mathcal{A}\bigg(\sum\limits_{j = m+1}^n\mathscr{H}(u_{j-1}\, )\bigg)\right\} +\mathcal{A}^{-1} \left\{\frac{1}{v^\alpha}\mathcal{A}\bigg(\sum\limits_{j = m+1}^n\mathscr{M}(u_{j-1}\, )\bigg)\right\} \\ +\mathcal{A}^{-1} \left\{\frac{1}{v^\alpha}\mathcal{A}\bigg(\sum\limits_{j = m+1}^n(A_{j-1}\, )\bigg)\right\}\Bigg| \\ = \max\limits_{t\in J}\Bigg|\mathcal{A}^{-1} \left\{\frac{1}{v^\alpha}\mathcal{A}\bigg(\sum\limits_{j = m}^{n-1}\mathscr{H}(u_{j}\, )\bigg)\right\} +\\ \mathcal{A}^{-1} \left\{\frac{1}{v^\alpha}\mathcal{A}\bigg(\sum\limits_{j = m}^{n-1}\mathscr{M}(u_{j}\, )\bigg)\right\} +\mathcal{A}^{-1} \left\{\frac{1}{v^\alpha}\mathcal{A}\bigg(\sum\limits_{j = m}^{n-1}(A_{j}\, )\bigg)\right\}\Bigg|\\ \leq\max\limits_{t\in J}\Bigg|\mathcal{A}^{-1} \left\{\frac{1}{v^\alpha}\mathcal{A}\bigg(\sum\limits_{j = m}^{n-1}\mathscr{H}(\mathscr{S}_{n-1})-\mathscr{H}(\mathscr{S}_{m-1})\bigg)\right\} \\+\mathcal{A}^{-1} \left\{\frac{1}{v^\alpha}\mathcal{A}\bigg(\sum\limits_{j = m}^{n-1}\mathscr{M}(\mathscr{S}_{n-1})-\mathscr{M}(\mathscr{S}_{m-1})\bigg)\right\} \\ +\mathcal{A}^{-1} \left\{\frac{1}{v^\alpha}\mathcal{A}\bigg(\sum\limits_{j = m}^{n-1}\mathscr{N}(\mathscr{S}_{n-1})-\mathscr{N}(\mathscr{S}_{m-1})\bigg)\right\}\Bigg| \end{array}$

$\begin{equation*} \begin{split} &\leq \ell_1\max\limits_{t\in J}\Bigg|\mathcal{A}^{-1} \left\{\frac{1}{v^\alpha}\mathcal{A}\bigg((\mathscr{S}_{n-1})-(\mathscr{S}_{m-1})\bigg)\right\}\Bigg| +\ell_2\max\limits_{t\in J}\Bigg|\mathcal{A}^{-1} \left\{\frac{1}{v^\alpha}\mathcal{A}\bigg((\mathscr{S}_{n-1})-(\mathscr{S}_{m-1})\bigg)\right\}\Bigg| \\&+\ell_3\max\limits_{t\in J}\Bigg|\mathcal{A}^{-1} \left\{\frac{1}{v^\alpha}\mathcal{A}\bigg((\mathscr{S}_{n-1})-(\mathscr{S}_{m-1})\bigg)\right\}\Bigg| = \frac{(\ell_1+\ell_2+\ell_3)t^{\alpha-1}}{\Gamma(\alpha)}\left\lVert{\mathscr{S}_{n-1}-\mathscr{S}_{m-1}}\right\rVert. \end{split} \end{equation*}$

Choosing $n = m+1$ then

$\left\lVert{\mathscr{S}_{m+1}-\mathscr{S}_{m}}\right\rVert\leq\gamma\left\lVert{\mathscr{S}_m-\mathscr{S}_{m-1}}\right\rVert\leq\gamma^2\left\lVert{\mathscr{S}_{m-1}-\mathscr{S}_{m-2}}\right\rVert\leq\cdots\cdots \leq\gamma^m\left\lVert{\mathscr{S}_1-\mathscr{S}_{0}}\right\rVert,$

with $\gamma = \frac{(\ell_1+\ell_2+\ell_3)t^{\alpha-1}}{\Gamma(\alpha)},$ by using the following triangular inequality

$\begin{equation*} \begin{split}\left\lVert{\mathscr{S}_n-\mathscr{S}_{m}}\right\rVert&\leq\left\lVert{\mathscr{S}_m-\mathscr{S}_{m+1}}\right\rVert\leq\left\lVert{\mathscr{S}_{m+1}-\mathscr{S}_{m+2}}\right\rVert\leq\cdots\cdots \leq\gamma^m\left\lVert{\mathscr{S}_1-\mathscr{S}_{0}}\right\rVert\\& \leq\left(\gamma^m+\gamma^{m+1}+\cdots+\gamma^n\right)\left\lVert{\mathscr{S}_1-\mathscr{S}_0}\right\rVert\\& \leq\gamma^m\left(1+\gamma+\gamma^2+\cdots+\gamma^{n-m-1}\right)\left\lVert{\mathscr{S}_1-\mathscr{S}_0}\right\rVert\\& \leq\gamma^m\left(\frac{1-\gamma^{n-m}}{1-\gamma}\right)\left\lVert{u_1}\right\rVert. \end{split} \end{equation*}$

Now by definition $0 < \gamma < 1$ , we have $1-\gamma^{n-m} < 1$ , thus we have

$\begin{equation} \left\lVert{\mathscr{S}_n-\mathscr{S}_m}\right\rVert\leq\frac{\gamma^m}{1-\gamma}\max\limits_{t\in J}\left\lVert{u_1}\right\rVert, \end{equation}$

(4.2)

and also as $|u| < \infty$ ( $u\,$ is bounded), therefore, $\left\lVert{\mathscr{S}_n-\mathscr{S}_m}\right\rVert\rightarrow 0$ , hence $\mathscr{S}_n$ is a Cauchy sequence in the Banach space $\mathscr{B}$ , hence $\sum_{j = 0}^{n}u_j\,$ is convergent.

5. Applications of ADM

In this section, we consider two specific examples of time-fractional Korteweg-de Vries (TF-KdV) and time-fractional modified Korteweg-de Vries (TF-mKdV) equations in the form of Eq (1.1) with some initial conditions and apply ADM to find their approximate solutions.

Next we consider two examples one for time-fractional order KdV and the other for time-fractional mKdV and use the propose method (ADM) to obtain an approximate solution.

Example 1. For $m = 1$ in Eq (3.1) we can take the following TF-KdV in Caputo's form ^[39]

$\begin{eqnarray} ^cD_t^\alpha u+au\frac{\partial u}{\partial x}+b\frac{\partial ^3u}{\partial x^3}& = &0, \ \ \, 0 < \alpha\leq 1, \end{eqnarray}$

(5.1)

with initial condition of the form

$\begin{eqnarray} u_0 = u(x, 0) = A \operatorname{sech}^2(kx). \end{eqnarray}$

(5.2)

The exact solution for $\alpha = 1$ of Eq (5.1) ^[39]

$\begin{equation} u\, = A\, \operatorname{sech}^2(kx-4k^3bt), \end{equation}$

(5.3)

where $A = \frac{12k^2b}{a}$ is amplitude and $4 b k^3$ is the speed of the wave function. Applying ADM to Eq (5.1) and decomposing the non-linear term $uu_x$ by using the Adomian formula (3.7), we obtain the following Adomian polynomials

$\begin{array}{l} A_0 = u_0u_{0x}, \quad A_1 = u_0u_{1x}+u_1u_{0x}, \quad A_2 = u_0u_{2x}+u_1u_{1x}+u_2u_{0x},\\ \quad A_3 = u_0u_{3x}+u_1u_{2x}+u_3u_{0x}+u_2u_{1x}, \end{array}$

putting the above values in Eq (3.8) with assumption $h_0(x) = u_0(x, 0)$ with $h_1(x) = h_2(x) = h_3(x) = 0$ , we obtain the following single soliton solution for the TF-KdV Eq (5.1) in the form

$\begin{eqnarray*} u_0 = h_0(x), \quad u_1 = \frac{t^\alpha}{\Gamma(\alpha+1)}f_1(x), \quad u_2 = -\frac{t^{2\alpha}}{\Gamma(2\alpha+1)}f_2(x), \quad u_3 = \frac{t^{3\alpha}}{\Gamma(3\alpha+1)}f_3(x), \end{eqnarray*}$

where

$\begin{eqnarray*} &f_1(x)& = ah_0h_{0x}+bh_{0xxx}, \quad f_2(x) = a(h_0f_{1x}+2h_0f_1h_{0x})+bf_{1xxx}, \\ &f_3(x)& = a\left[h_0f_{2x}+2h_0f_2h_{0x}+\frac{\Gamma(2\alpha+1)}{\Gamma(\alpha+1)}(2h_0f_1f_{1x}h_{0x}+f_{1}h_{0x})\right]+bf_{2xxx}. \end{eqnarray*}$

The final solution in the series form up to $\mathcal{O}(4)$ is given by

$\begin{equation} U(x, t) = h_0(x)+\frac{t^\alpha}{\Gamma(\alpha+1)}f_1(x)-\frac{t^{2\alpha}}{\Gamma(2\alpha+1)}f_2(x)+ \frac{t^{3\alpha}}{\Gamma(3\alpha+1)}f_3(x). \end{equation}$

(5.4)

For $\alpha = 1$ , we obtain

$\begin{equation*} u(x, t)\, = \frac{12k^2b}{a}\, \operatorname{sech}^2(kx-4k^3bt). \end{equation*}$

The ADM solution for different $\alpha$ are reported in , by considering $k = 0.1, a = 2, b = 0.2, t = 1$ .

Table 1. Evaluation between approximate (

$U(x, t))$ versus an exact solution (u(x, t) for different values a = 0.2, b = 0.4, k = 0.1,

$-10\leq x\leq 10$ and

$0\leq t\leq 1$ Exp.1.

t	$U(\alpha=0.3)$	$U(\alpha=0.5)$	$U(\alpha=0.7)$	$U(\alpha=0.9)$	$U(\alpha=1)$	Exact	MAE
x=-10
0.00	0.0050869	0.0050869	0.0050869	0.0050869	0.0050869	0.0050869
0.25	0.0050056	0.0050276	0.0050447	0.0050575	0.0050625	0.0050634
0.50	0.0049834	0.0049994	0.0050155	0.0050302	0.0050367	0.0050400	4.0619E-05
0.75	0.0049673	0.0049763	0.0049888	0.0050025	0.0050093	0.0050167
1.00	0.0049542	0.0049559	0.0049633	0.0049742	0.0049803	0.0049935
x=-5
0.00	0.030238	0.030238	0.030238	0.030238	0.030238	0.030238
0.25	0.029905	0.029984	0.030052	0.030105	0.030127	0.030128
0.50	0.029825	0.029877	0.029933	0.029989	0.030015	0.030018	1.2525E-04
0.75	0.02977	0.029793	0.029831	0.0050025	0.029902	0.029908
1.00	0.029726	0.029722	0.029737	0.0049742	0.029787	0.029798
x=0
0.00	0.0720000	0.0720000	0.0720000	0.0720000	0.0720000	0.0720000
0.25	0.071994	0.071997	0.071998	0.071999	0.0720000	0.0720000
0.50	0.07199	0.071993	0.071996	0.071998	0.0719987	0.0719987	6.2213E-06
0.75	0.071988	0.07199	0.071993	0.071995	0.071996	0.071996
1.00	0.071985	0.071987	0.071989	0.071992	0.071993	0.071993
x=5
0.00	0.030238	0.030238	0.030238	0.030238	0.030238	0.030238
0.25	0.030555	0.030483	0.03042	0.030369	0.030348	0.030349
0.50	0.030625	0.030582	00.030532	0.030481	0.030457	0.03046	1.0040E-04
0.75	0.030673	0.030657	0.030626	0.030587	0.030565	0.030571
1.00	0.030711	0.030719	0.03071	0.030687	0.030671	0.030682
x=10
0.00	0.0050869	0.0050869	0.0050869	0.0050869	0.0050869	0.0050869
0.25	0.005144	0.0051338	0.0050447	0.0051233	0.0051096	0.0051105
0.50	0.0051538	0.0051496	0.0050155	0.0051431	0.0051308	0.0051341	1.4575E-05
0.75	0.0051598	0.0051603	0.0049888	0.0051583	0.0051505	0.0051579
1.00	0.0051641	0.0051684	0.0049633	0.0051706	0.0051687	0.0051819

| Show Table

DownLoad: CSV

, displays the zeroth and higher order solitary wave solutions (5.3) and (5.4) versus the space variable $x$ respectively for time-fractional KdV Eq (5.1) for small time of $t$ . It infers that at $t > 0$ the higher order solitary wave solution is significantly modified. The wave three-dimensional profiles for solutions (5.3) and (5.4) in , also reveal that variations in temporal variables affect the amplitude and spatial extensions of the higher order wave solutions. We have plotted the approximate solution (5.3) against $t$ at $x = 0.5$ and $1$ respectively, with variation in the time-fractional index $\alpha$ [see , ]. A degree of enhancement in $\alpha$ reduces the spatial width of the solitary waves. Thus, the time-fractional index leads to a reduction in the wave dispersion that in turn localises the wave profile. The effect of the non-linearity coefficient "a" is shown in . We see that the amplitude of the solitary wave decreases regularly when the values of the non-linearity coefficient are increasing. This definitely occurs for wave equations because the non-linearity coefficient has an inverse proportionality with the solitary wave amplitude. On the other hand, the wave amplitude is gradually increasing with the increasing values of the dispersion coefficient "b" in . The dispersive coefficient is proportional to the amplitude of a solitary wave. We have depicted the higher order solution (5.4) for the TF-KdV Eq (5.1), verses $x$ at time $t = 3$ (solid curve), at 3.5 (dashed curve), and at 4 (dotted curve) [see ]. Obviously, the enhancement in $t$ gives rise to the pulse amplitude while increasing the spatial width. Moreover, the pulse shape solutions suffer from oscillations for $t > 0$ . The wave solution (5.4) presented in at different values of the temporal index ( $\alpha$ ) reveals strengthening of the wave with wider spatial extension at large $\alpha$ . One can see that the solitary wave solutions for the TF-KdV equation are observed. By employing an Aboodh decomposition (ADM) technique and with the addition of $t$ the Caputo operator for the solitary excitations are derived in Eq (5.4). Importantly, the wave solution significantly changes in time. The large amplitude electrostatic excitations in the auroral zone ^[27], associated with the intensified electric field. These may be described by the solutions (5.3) and (5.4) for the TF-KdV equation of Exp.1. Our results obtained in this paper are so important because the waves produced by the solutions of the TF-KdV may have very high amplitudes. The amplitudes of these waves can be reduced by using the fractional order of the wave equations. In other words, by taking the fractional order of the time derivative term, a better analysis of the waves can be achieved.

Figure 1. In the top panel, (a) shows exact solution of Eq (5.3), while (b) shows comparison of approximate (ADM) and exact solution Eq (5.4)

$k = 0.1, a = 0.2, b = 0.4$ , for small time

$t = 0.2$ . The bottom panel (c, d) are corresponding surface 3D plots for both solutions against the temporal variables

$(x, t)$ with parameters (

$\alpha = 1$ for exact, and

$0.7$ , for the approximate solution, with fixed

$k = 0.1, a = 2, b = 0.2, t = 1$ ).

DownLoad: Full-Size Img PowerPoint

Figure 2. The above plots show the dispersive behaviours of the solitary waves by varying time and the fractional order

$\alpha$ with fixed special variable (a)

$x = 0.5$ and (b)

$x = 1$ of the exact (

$\alpha = 1$ star black) solution Eq (5.3) and approximate (

$\alpha < 1$ blue) solution Eq (5.4) respectively.

DownLoad: Full-Size Img PowerPoint

Figure 3. Solitary waves profiles for different values of (a) the non-linearity coefficient "a" with b = 0.4 and (b) the dispersive coefficient "b" with a = 0.2, when the wave number k = 0.1, the fractional index

$\alpha = 0.5$ and t = 0.5 of Eq (5.4).

DownLoad: Full-Size Img PowerPoint

Figure 4. Plots of the approximate solution for (a) different values of t with

$\alpha = 0.9$ (b) different values of

$\alpha$ for larger time and the other parameters are

$a = 0.9, k = 0.5, b = 0.9$ of Eq (5.4).

DownLoad: Full-Size Img PowerPoint

Example 2. For $m = 2, a = 6, b = 1$ , the modified time-fractional KdV ^[26] in Caputo's sense is given by

$\begin{eqnarray} &^CD_t^\alpha\, v+6\, v^2\frac{\partial\, v}{\partial x}+\frac{\partial ^3\, v}{\partial x^3} = 0, \, \, \, 0 < \alpha\leq 1, \end{eqnarray}$

(5.5)

with

$\begin{eqnarray} v_0 = v(x, 0) = \sqrt{c} \operatorname{sech}^2(k+\sqrt{c}x). \end{eqnarray}$

(5.6)

When $\alpha = 1$ , Eq (5.5) gives ^[26]

$\begin{equation} v(x, t) = \sqrt{c}\, \operatorname{sech}(k+\sqrt{c}x-ct). \end{equation}$

(5.7)

Applying the ADM to Eq (5.5) and decompose the non-linear term $v^2v_x$ in Eq (5.5) by using the Adomian formula (3.7), we obtain

$\begin{array}{l} A_0 = v_0^2v_{0x}, \quad A_1 = v_0^2v_{1x}+v_1^2v_{0x}, \quad A_2 = v_0^2v_{2x}+v_1^2v_{1x}+v_2^2v_{0x}, \\ A_3 = v_0^2v_{3x}+v_1^2v_{2x}+v_3^2v_{0x}+v_2^2v_{1x}. \end{array}$

Putting $a = 6$ and $b = 1$ in Eq (3.8) and assume that $g_0(x) = v(x, 0)$ with $g_1(x) = g_2(x) = g_3(x)$ , we obtain

$\begin{eqnarray*} v_0 = g_0(x), \quad v_1 = \frac{t^\alpha}{\Gamma(\alpha+1)}q_1(x), \quad v_2 = -\frac{t^{2\alpha}}{\Gamma(2\alpha+1)}q_2(x), \quad v_3 = \frac{t^{3\alpha}}{\Gamma(3\alpha+1)}q_3(x), \end{eqnarray*}$

where

$\begin{eqnarray*} &q_1(x)& = 6g_0^2g_{0x}+g_{0xxx}, \quad q_2(x) = 6(g_0^2q_{1x}+2g_0q_1g_{0x})+q_{1xxx}, \\ &q_3(x)& = 6\left[g_0^2q_{2x}+2g_0q_2g_{0x}+\frac{\Gamma(2\alpha+1)}{\Gamma(\alpha+1)}(2g_0q_1q_{1x}g_{0x}+q_{1}^2g_{0x})\right]+q_{2xxx}. \end{eqnarray*}$

The solution up to $\mathcal{O}(4)$ can be written as

$\begin{equation} V(x, t) = g_0(x)+\frac{t^\alpha}{\Gamma(\alpha+1)}q_1(x)-\frac{t^{2\alpha}}{\Gamma(2\alpha+1)}q_2(x)+ \frac{t^{3\alpha}}{\Gamma(3\alpha+1)}q_3(x). \end{equation}$

(5.8)

For $\alpha = 1$ we found the following exact solution

$\begin{equation*} v(x, t) = \sqrt{c}\, \operatorname{sech}(k+\sqrt{c}x-ct). \end{equation*}$

We have displayed the zeroth (higher) order solitary wave solutions (5.7) and (5.8) in , respectively, for the modified time-fractional KdV Eq (5.5) verses x for small t. Notice that the wave solution of the TF-mKdV equation suffers only a relatively modest steeping effect with a reduction in the wave dispersion. Moreover, at $t > 0$ the higher-order solution illustrates noticeable modifications in the wave profiles. The three dimensional profiles Figure 5(c, d) for the time fractional (exact and approximate) solutions also confirm that both the solutions are quite similar. , display the solution (5.8) at $x = 0.5$ and 1 respectively with variation in $\alpha$ . Contrary to the TF-KdV solution, the wave solution for TF-mKdV admits reduced wave dispersion at $t > 0$ . We observed that, the dispersive approximate solution ( $\alpha < 1$ the blue curves) approaching to the exact solution $(\alpha = 1$ , black curve) when then $\alpha$ values are approaching to 1. In Figure 7(a), we depicted the approximate (ADM) solution against the special variable x by changing the wave speed "c = 0.1 (solid curve), 0.2 (dashed curve), and 0.3 (dotted curve)". We observed that when the speed "c" is increasing, the wave amplitude is increasing while its width is decreasing, which shows that the amplitude is directly proportional to the wave speed while the width of the wave is inversely related to the wave speed. Similarly, in , we displayed the approximate solution against "x" by changing the wave number "k = 0.5 (solid curve), 1 (dashed curve), and 1.5 (dotted curve)". We see that, for different values of the wave number and equal amplitudes, wave profiles are obtained. The wave solution (5.5) for TF-mKdV in shows at different times that the solitary waves suffer from oscillation due to the external perturbations. It also shows that the solitary solution for the TF-mKdV is relatively less stable against the perturbations. Moreover, the increase in $t$ and $\alpha$ modify the amplitude and width of the wave. Variation of the $\alpha$ value for the approximate solution is shown in against $x$ . We observed from the , when the temporal variable (t) and the fractional order $\alpha$ increasing the amplitudes of the wave profile getting larger values.

Figure 5. In the top panel, (a) shows exact solution for Eq (5.7) while (b) shows a comparison of exact and approximate solution (ADM) of Eq (5.8). The bottom panel (c, d) are the corresponding surface 3D plots against the temporal variables

$(x, t)$ with parameters (

$\alpha = 1$ for exact, and

$0.7$ , for the approximate solution, with fixed

$k = 0, c = 1, t = 0.2$ ).

DownLoad: Full-Size Img PowerPoint

Figure 6. Dispersive behaviours of the solitary waves by varying time and the fractional order

$\alpha$ with fixed special variable (a)

$x = 0.5$ and (b)

$x = 1$ for exact (

$\alpha = 1$ stars blue curve of Eq (5.7)) and approximate (

$\alpha < 1$ blue curves Eq (5.8)) respectively.

DownLoad: Full-Size Img PowerPoint

Figure 7. Solitary waves profiles for different values of (a) the speed "c" with k = 0.5 and (b) the wave number "k", with c = 0.2 when the fractional index

$\alpha = 0.7$ and t = 0.5 of Eq (5.8).

DownLoad: Full-Size Img PowerPoint

Figure 8. Plots of the approximate solution for (a) different values of

$t$ with

$\alpha = 0.9$ (b) different values of

$\alpha$ for larger time values and other parameters are

$c = 0.6, k = 0$ of Eq (5.8).

DownLoad: Full-Size Img PowerPoint

The mean absolute error (MAE) is a statistic that calculates the variance in errors amongst matching interpretations representing identical singularities. The relationships between an exact and an approximate solution, future time against beginning time, and one measuring method versus another quantity technique are all examples of $y$ versus $x$ . The MAE can be computed as

$\begin{eqnarray} MAE = \frac{1}{n}\sum\limits_{i = 1}^{n}|y_i-x_i|, \end{eqnarray}$

(5.9)

where, $x_i$ is the expected value and $y_i$ is the real value of the system. The following figures are obtained from the error analysis in the Tables 1–4.

Table 2. Comparison between the approximate solution (

$V(x, t)$ ), versus an exact solution (

$v(x, t)$ ) for

$-10\leq x\leq 10$ and

$0\leq t\leq 1$ with k = 0.5 and c = 0.82 of Exp.2.

t	$V(\alpha=0.3)$	$V(\alpha=0.5)$	$V(\alpha=0.7)$	$V(\alpha=0.9)$	$V(\alpha=1)$	Exact	MAE
x=-10
0.00	0.010216	0.010216	0.010216	0.010216	0.010216	0.010216
0.25	0.0095819	0.00972	0.0098439	0.0099468	0.0099898	0.0097175
0.50	0.0094455	0.0095258	0.0096209	0.0097203	0.0097689	0.0092437	3.6330E-04
0.75	0.0093537	0.0093809	0.0094359	0.0095107	0.0095529	0.008793
1.00	0.0092825	0.0092616	0.0092729	0.0093128	0.0093417	0.0083642
x=-5
0.00	0.094515	0.094515	0.094515	0.094515	0.094515	0.094515
0.25	0.088823	0.090001	0.091122	0.092066	0.09246	0.090002
0.50	0.087702	0.088282	0.089091	0.08999	0.090436	0.085696	3.5E-03
0.75	0.086993	0.087053	0.087436	0.08807	0.088455	0.081589
1.00	0.086476	0.08609	0.086022	0.086289	0.08653	0.077671
x=0
0.00	0.447210	0.447210	0.447210	0.447210	0.447210	0.447210
0.25	0.445160	0.446220	0.446770	0.447030	0.44710	0.446660
0.50	0.444000	0.445150	0.446010	0.446560	0.446740	0.444990	1.46E-03
0.75	0.443020	0.444030	0.445030	0.445830	0.446130	0.442230
1.00	0.442140	0.442860	0.443860	0.444830	0.445250	0.438420
x=5
0.00	0.094515	0.094515	0.094515	0.094515	0.094515	0.094515
0.25	0.099243	0.099243	0.099243	0.099243	0.099243	0.030349
0.50	0.104190	0.104190	0.10419	0.030481	0.10419	0.104190	4.446E-03
0.75	0.109380	0.109380	0.109380	0.0305870	0.10938	0.109380
1.00	0.114800	0.114800	0.114800	0.0306870	0.1148	0.030682
x=10
0.00	0.010216	0.010216	0.010216	0.010216	0.010216	0.010216
0.25	0.010928	0.010752	0.010606	0.010492	0.010447	0.010739
0.50	0.011106	0.01117	0.011083	0.010978	0.010683	0.01129	4.3146E-04
0.75	0.01133	0.011332	0.011289	0.011215	0.010924	0.011868
1.00	0.114800	0.11480	0.114800	0.030687	0.011171	0.012477

| Show Table

DownLoad: CSV

Table 3. Comparison of exact solution (u(x, t)) with approximate (U(x, t)) solution for

$\alpha = 0.5$ , a = 0.2, b = 0.4, and k = 1 of Exp.1.

$x$	t	$\xi$ = $kx-4bk^3t$	Exact solution	ADM Solution	Error
-50	0.00	-49.9434	-0.35995	-0.35995	2.5009 $\times 10^{-08}$
-40	0.25	-39.6454	-0.35965	-0.35965	1.9820 $\times 10^{-07}$
-30	0.50	-29.6453	-0.35745	-0.35746	2.1868 $\times 10^{-06}$
-20	0.75	-19.6423	-0.34165	-0.34170	5.1926 $\times 10^{-05}$
$\cdots$	$\cdots$	$\cdots$	$\cdots$	$\cdots$	$\cdots$
$\cdots$	$\cdots$	$\cdots$	$\cdots$	$\cdots$	$\cdots$
$\cdots$	$\cdots$	$\cdots$	$\cdots$	$\cdots$	$\cdots$
$\cdots$	$\cdots$	$\cdots$	$\cdots$	$\cdots$	$\cdots$
$\cdots$	$\cdots$	$\cdots$	$\cdots$	$\cdots$	$\cdots$
20	0.25	20.3634	0.11987	0.11988	-1.6363 $\times 10^{-05}$
30	0.50	30.3653	0.12000	0.12000	-3.3767 $\times 10^{-07}$
40	0.75	40.3634	0.12000	0.12000	-6.2851 $\times 10^{-09}$
50	1.00	50.3643	0.12000	0.12000	-1.1537 $\times 10^{-10}$

| Show Table

DownLoad: CSV

Table 4. Comparison of exact (v(x, t)) solution with approximate (V(x, t) solution for k = 0.5, and c = 0.82 of Exp.2.

$x$	t	$\xi$ = $k+\sqrt{c}x-ct$	Exact solution	ADM Solution	Error
-50	0.00	-45.5969	-0.35995	-0.35995	-2.8783 $\times 10^{-07}$
-40	0.25	-36.5415	-0.35966	-0.35966	-2.1145 $\times 10^{-06}$
-30	0.50	-27.4862	-0.35751	-0.35750	-1.4960 $\times 10^{-05}$
-20	0.75	-18.4308	-0.34207	-0.34199	-7.7286 $\times 10^{-05}$
$\cdots$	$\cdots$	$\cdots$	$\cdots$	$\cdots$	$\cdots$
$\cdots$	$\cdots$	$\cdots$	$\cdots$	$\cdots$	$\cdots$
$\cdots$	$\cdots$	$\cdots$	$\cdots$	$\cdots$	$\cdots$
$\cdots$	$\cdots$	$\cdots$	$\cdots$	$\cdots$	$\cdots$
$\cdots$	$\cdots$	$\cdots$	$\cdots$	$\cdots$	$\cdots$
20	0.25	17.7908	0.11986	0.11987	-1.3770 $\times 10^{-05}$
30	0.50	26.8462	0.12000	0.12000	-2.8188 $\times 10^{-07}$
40	0.75	35.9015	0.12000	0.12000	-5.2413 $\times 10^{-09}$
50	1.00	44.9569	0.12000	0.12000	-9.6194 $\times 10^{-11}$

| Show Table

DownLoad: CSV

The accuracy and effectiveness of the ADM are confirmed by using the error analysis given in Tables 1 and 2 followed by its MAE Figure 9(a, b) respectively. The other two Tables 3 and 4 confirm the errors that occur between the exact and approximate solution of the model. Their plots Figure 10(a, b) are shown, which are the graphical conformations for the error analysis between the two solutions (exact and approximate).

Figure 9. (a, b) 3D plots represent the mean absolute error (MAE) for different values of x and t for Tables 1 and 2 respectively.

DownLoad: Full-Size Img PowerPoint

Figure 10. (a, b) 3D plots represent the mean absolute error (MAE) for different values of x and t for Tables 3 and 4 respectively.

DownLoad: Full-Size Img PowerPoint

6. Conclusions

We consider time-fractional Korteweg-de Vries (TF-KdV) and time-fractional modified Korteweg-de Vries (TF-mKdV) equations for solitary wave solutions with the help of an analytical method called the "Aboodh decomposition method (ADM)." The proposed method is a combination of the Aboodh transform and the decomposition method, which is a very authentic and valuable method for solving fractional-order non-linear problems. The approximate solution obtained with the help of ADM is compared with the exact solution of the models to verify the efficiency of the method. We observed from the graphical analysis in the manuscript the time-fractional order $\alpha$ can modify the wave profiles in such a way that when $\alpha$ values is increasing the amplitudes of the wave is reducing which is very much important to study the small amplitude's characteristics of such wave equations. The effect of the nonlinearity coefficient "a" and the dispersive coefficient "b" in Exp.1 are also discussed, form the graphical analysis we have shown that both of these coefficients have significant affect on the wave amplitudes as well as its widths. We have studied the higher-order series solutions (up to fourth order) for time-fractional KdV and modified KdV equations with the help of the ADM. The convergence analysis determined that the obtained results are similar to the exact solutions of the models. The uniqueness results confirmed that ADM is an effective and systematic scheme for solving non-linear dynamical problems. The analytical and numerical solutions are plotted by marginal changes in parameters $x$ , $t$ and $\alpha$ . It is observed that a significant variation in the index values causes a large variation in the amplitude of the wave profile. For small-time $t$ , the series solutions coincide with the exact solutions of the models, and the solutions are stable. However, for large-time $t$ the results show that the amplitudes of the wave profiles no longer remain constant at the given parameters. The results are tabulated for altered $x$ , $t$ and $\alpha$ . It is observed that mean absolute error (MAE) is decreasing for small-time and large values of $\alpha$ (fractional order). This means that for taking $\alpha$ values nearest to 1 (exact), the MAE values decrease, which means that the obtained results coincide with the solution of the considered models.

7. Future work

We studied the analytical solution of the TF-KdV and TF-mKdV equations with the help of ADM. The effectiveness of the proposed method is studied with the help of graphical analysis and some numerical error tables. The method is also essential for $K(m, n)$ equation obtained in ^[40]. Here, the author shows that $K(m, n)$ equations are only Lax integrable for specific values of the parameters $(\alpha, m, n)$ . For the integrable instances, nontrivial prolongation structures and Lax pairings are provided.

Acknowledgments

Princess Nourah bint Abdulrahman University Researchers Supporting Project number (PNURSP2022R8). Princess Nourah bint Abdulrahman University, Riyadh, Saudi Arabia.

Conflict of interest

The authors declare that they have no conflicts of interest.

References

[1]	R. Nigam, R. Tripathy, S. Dutta, N. Bhagia, R. Nagori, K. Chandrasekar, et al., Crop type discrimination and health assessment using hyperspectral imaging, Curr. Sci., 116 (2019), 1108–1123. https://www.jstor.org/stable/27138003
[2]	J. ten Harkel, H. Bartholomeus, L. Kooistra, Biomass and crop height estimation of different crops using UAV-based LiDAR, Remote Sens., 12 (2020), 17. https://doi.org/10.3390/rs12010017 doi: 10.3390/rs12010017
[3]	U. S. Panday, N. Shrestha, S. Maharjan, A. K. Pratihast, Shahnawaz, K. L. Shrestha, et al., Correlating the plant height of wheat with above-ground biomass and crop yield using drone imagery and crop surface model, a case study from nepal, Drones, 4 (2020), 28. https://doi.org/10.3390/drones4030028 doi: 10.3390/drones4030028
[4]	A. Michez, P. Lejeune, S. Bauwens, A. A. L. Herinaina, Y. Blaise, E. C. Muñoz, et al., Mapping and monitoring of biomass and grazing in pasture with an unmanned aerial system, Remote Sens., 11 (2019), 473. https://doi.org/10.3390/rs11050473 doi: 10.3390/rs11050473
[5]	A. I. de Castro, R. Ehsani, R. C. Ploetz, J. H. Crane, S. Buchanon, Detection of laurel wilt disease in avocado using low altitude aerial imaging, PloS ONE, 10 (2015), 1–13. https://doi.org/10.1371/journal.pone.0124642 doi: 10.1371/journal.pone.0124642
[6]	A. Mahlein, Plant disease detection by imaging sensors–parallels and specific demands for precision agriculture and plant phenotyping, Plant Dis., 100 (2016), 241–251. https://doi.org/10.1094/PDIS-03-15-0340-FE doi: 10.1094/PDIS-03-15-0340-FE
[7]	P. Moghadam, D. Ward, E. Goan, S. Jayawardena, P. Sikka, E. Hernandez, Plant disease detection using hyperspectral imaging, in 2017 International Conference on Digital Image Computing: Techniques and Applications (DICTA), (2017), 1–8. https://doi.org/10.1109/DICTA.2017.8227476
[8]	D. Gómez-Candón, J. Torres-Sanchez, S. Labbé, A. Jolivot, S. Martinez, J. L. Regnard, Water stress assessment at tree scale: high-resolution thermal UAV imagery acquisition and processing, Acta Hortic., 1150 (2017), 159–166. https://doi.org/10.17660/ActaHortic.2017.1150.23 doi: 10.17660/ActaHortic.2017.1150.23
[9]	C. A. Reynolds, M. Yitayew, D. C. Slack, C. F. Hutchinson, A. Huete, M. S. Petersen, Estimating crop yields and production by integrating the FAO Crop Specific Water Balance model with real-time satellite data and ground-based ancillary data, Int. J. Remote Sens., 21 (2000), 3487–3508. https://doi.org/10.1080/014311600750037516 doi: 10.1080/014311600750037516
[10]	S. S. Panda, D. P. Ames, S. Panigrahi, Application of vegetation indices for agricultural crop yield prediction using neural network techniques, Remote Sens., 2 (2010), 673–696. https://doi.org/10.3390/rs2030673 doi: 10.3390/rs2030673
[11]	Z. Fu, J. Jiang, Y. Gao, B. Krienke, M. Wang, K. Zhong, et al., Wheat growth monitoring and yield estimation based on multi-rotor unmanned aerial vehicle, Remote Sens., 12 (2020), 508. https://doi.org/10.3390/rs12030508 doi: 10.3390/rs12030508
[12]	S. Guan, K. Fukami, H. Matsunaka, M. Okami, R. Tanaka, H. Nakano, et al., Assessing correlation of high-resolution NDVI with fertilizer application level and yield of rice and wheat crops using small UAVs, Remote Sens., 11 (2019), 112. https://doi.org/10.3390/rs11020112 doi: 10.3390/rs11020112
[13]	M. Maimaitijiang, V. Sagan, P. Sidike, S. Hartling, F. Esposito, F. B. Fritschi, Soybean yield prediction from UAV using multimodal data fusion and deep learning, Remote Sens. Environ., 237 (2020), 111599. https://doi.org/10.1016/j.rse.2019.111599 doi: 10.1016/j.rse.2019.111599
[14]	Y. Guo, S. Chen, X. Li, M. Cunha, S. Jayavelu, D. Cammarano, et al., Machine learning-based approaches for predicting SPAD values of maize using multi-spectral images, Remote Sens., 14 (2022), 1337. https://doi.org/10.3390/rs14061337 doi: 10.3390/rs14061337
[15]	Z. Sun, X. Wang, Z. Wang, L. Yang, Y. Xie, Y. Huang, UAVs as remote sensing platforms in plant ecology: review of applications and challenges, J. Plant Ecol., 14 (2021), 1003–1023. https://doi.org/10.1093/jpe/rtab089 doi: 10.1093/jpe/rtab089
[16]	J. Xue, B. Su, Significant remote sensing vegetation indices: A review of developments and applications, J. Sens., 2017 (2017), 1–17. https://doi.org/10.1155/2017/1353691 doi: 10.1155/2017/1353691
[17]	L. Wan, H. Cen, J. Zhu, J. Zhang, Y. Zhu, D. Sun, et al., Grain yield prediction of rice using multi-temporal UAV-based RGB and multispectral images and model transfer-a case study of small farmlands in the South of China, Agric. For. Meteorol., 291 (2020), 108096. https://doi.org/10.1016/j.agrformet.2020.108096 doi: 10.1016/j.agrformet.2020.108096
[18]	J. Zhou, J. Zhou, H. Ye, M. L. Ali, P. Chen, H. T. Nguyen, Yield estimation of soybean breeding lines under drought stress using unmanned aerial vehicle-based imagery and convolutional neural network, Biosyst. Eng., 204 (2021), 90–103. https://doi.org/10.1016/j.biosystemseng.2021.01.017 doi: 10.1016/j.biosystemseng.2021.01.017
[19]	Y. Guo, H. Wang, Z. Wu, S. Wang, H. Sun, J. Senthilnath, et al., Modified red blue vegetation index for chlorophyll estimation and yield prediction of maize from visible images captured by UAV, Sensors, 20 (2020), 5055. https://doi.org/10.3390/s20185055 doi: 10.3390/s20185055
[20]	Y. Guo, Y. Fu, F. Hao, X. Zhang, W. Wu, X. Jin, et al., Integrated phenology and climate in rice yields prediction using machine learning methods, Ecol. Indic., 120 (2021), 106935. https://doi.org/10.1016/j.ecolind.2020.106935 doi: 10.1016/j.ecolind.2020.106935
[21]	Peanut company of Australia, How peanuts are grown, 2023. Available from: https://pca.com.au/pca-profile/how-peanuts-are-grown/
[22]	Z. Ji, Y. Pan, X. Zhu, D. Zhang, J. Wang, A generalized model to predict large-scale crop yields integrating satellite-based vegetation index time series and phenology metrics, Ecol. Indic., 137 (2022), 108759. https://doi.org/10.1016/j.ecolind.2022.108759 doi: 10.1016/j.ecolind.2022.108759
[23]	H. García-Martínez, H. Flores-Magdaleno, R. Ascencio-Hernández, A. Khalil-Gardezi, L. Tijerina-Chávez, O. R. Mancilla-Villa, et al., Corn grain yield estimation from vegetation indices, canopy cover, plant density, and a neural network using multispectral and RGB images acquired with unmanned aerial vehicles, Agriculture, 10 (2020), 277. https://doi.org/10.3390/agriculture10070277 doi: 10.3390/agriculture10070277
[24]	X. Zhou, H. B. Zheng, X. Q. Xu, J. Y. He, X. K. Ge, X. Yao, et al., Predicting grain yield in rice using multi-temporal vegetation indices from UAV-based multispectral and digital imagery, ISPRS J. Photogramm. Remote Sens., 130 (2017), 246–255. https://doi.org/10.1016/j.isprsjprs.2017.05.003 doi: 10.1016/j.isprsjprs.2017.05.003
[25]	D. C. Tsouros, S. Bibi, P. G. Sarigiannidis, A review on UAV-based applications for precision agriculture, Information, 10 (2019), 349. https://doi.org/10.3390/info10110349 doi: 10.3390/info10110349
[26]	J. Kim, S. Kim, C. Ju, H. Il Son, Unmanned aerial vehicles in agriculture: A review of perspective of platform, control, and applications, IEEE Access, 7 (2019), 105100–105115. https://doi.org/10.1109/ACCESS.2019.2932119 doi: 10.1109/ACCESS.2019.2932119
[27]	T. B. Shahi, C. Xu, A. Neupane, W. Guo, Machine learning methods for precision agriculture with UAV imagery: A review, Electron. Res. Arch., 30 (2022), 4277–4317. https://doi.org/10.3934/era.2022218 doi: 10.3934/era.2022218
[28]	A. P. M. Ramos, L. P. Osco, D. E. G. Furuya, W. N. Gonçalves, D. C. Santana, L. P. R. Teodoro, et al., A random forest ranking approach to predict yield in maize with uav-based vegetation spectral indices, Comput. Electron. Agric., 178 (2020), 105791. https://doi.org/10.1016/j.compag.2020.105791 doi: 10.1016/j.compag.2020.105791
[29]	J. Geipel, J. Link, W. Claupein, Combined spectral and spatial modeling of corn yield based on aerial images and crop surface models acquired with an unmanned aircraft system, Remote sens., 6 (2014), 10335–10355. https://doi.org/10.3390/rs61110335 doi: 10.3390/rs61110335
[30]	A. Ashapure, S. Oh, T. G. Marconi, A. Chang, J. Jung, J. Landivar, et al., Unmanned aerial system based tomato yield estimation using machine learning, in Autonomous Air and Ground Sensing Systems for Agricultural Optimization and Phenotyping IV, (2019). https://doi.org/10.1117/12.2519129
[31]	A. Matese, S. F. Di Gennaro, Beyond the traditional NDVI index as a key factor to mainstream the use of UAV in precision viticulture, Sci. Rep., 11 (2021), 1–13. https://doi.org/10.1038/s41598-021-81652-3 doi: 10.1038/s41598-021-81652-3
[32]	C. Bian, H. Shi, S. Wu, K. Zhang, M. Wei, Y. Zhao, et al., Prediction of field-scale wheat yield using machine learning method and multi-spectral UAV Ddata, Remote Sens., 14 (2022), 1474. https://doi.org/10.3390/rs14061474 doi: 10.3390/rs14061474
[33]	S. Fei, M. A. Hassan, Y. Xiao, X. Su, Z. Chen, Q. Cheng, et al., UAV-based multi-sensor data fusion and machine learning algorithm for yield prediction in wheat, Precision Agric., 24 (2022), 1–26. https://doi.org/10.1007/s11119-022-09938-8 doi: 10.1007/s11119-022-09938-8
[34]	A. Patrick, S. Pelham, A. Culbreath, C. C. Holbrook, I. J. De Godoy, C. Li, High throughput phenotyping of tomato spot wilt disease in peanuts using unmanned aerial systems and multispectral imaging, IEEE Instrum. Meas. Mag., 20 (2017), 4–12. https://doi.org/10.1109/MIM.2017.7951684 doi: 10.1109/MIM.2017.7951684
[35]	QGIS development team, QGIS Geographic Information System, 2023. Available from: https://www.qgis.org
[36]	F. I. Matias, M. V. Caraza-Harter, J. B. Endelman, FIELDimageR: An R package to analyze orthomosaic images from agricultural field trials, Plant Phenom. J., 3 (2020), 20005. https://doi.org/10.1002/ppj2.20005 doi: 10.1002/ppj2.20005
[37]	A. J. Smola, B. Schö lkopf, A tutorial on support vector regression, Satistics Comput., 14 (2004), 199–222. https://doi.org/10.1023/B:STCO.0000035301.49549.88 doi: 10.1023/B:STCO.0000035301.49549.88
[38]	X. Zeng, S. Yuan, Y. Li, Q. Zou, Decision tree classification model for popularity forecast of Chinese colleges, J. Appl. Math., (2014), 1–7. https://doi.org/10.1155/2014/675806 doi: 10.1155/2014/675806
[39]	L. Breiman, Random forests, Mach. Learn., 45 (2001), 5–32. https://doi.org/10.1023/A:1010933404324 doi: 10.1023/A:1010933404324
[40]	F. Pedregosa, G. Varoquaux, A. Gramfort, V. Michel, B. Thirion, O. Grisel, et al., Scikit-learn: Machine learning in Python, J. Mach. Learn. Res., 12 (2011), 2825–2830.
[41]	T. Hastie, S. Rosset, J. Zhu, H. Zou, Multi-class adaboost, Stat. Interface, 2 (2009), 349–360. https://doi.org/10.4310/SⅡ.2009.v2.n3.a8 doi: 10.4310/SⅡ.2009.v2.n3.a8
[42]	T. Chen, C. Guestrin, Xgboost: A scalable tree boosting system, in Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, (2016), 785–794. https://doi.org/10.1145/2939672.2939785
[43]	M. M. Li, W. Guo, B. Verma, K. Tickle, J. O'Connor, Intelligent methods for solving inverse problems of backscattering spectra with noise: a comparison between neural networks and simulated annealing, Neural Comput. Appl., 18 (2009), 423–430. https://doi.org/10.1007/s00521-008-0219-x doi: 10.1007/s00521-008-0219-x
[44]	C. Tsai, C. Hsia, S. Yang, S. Liu, Z. Fang, Optimizing hyperparameters of deep learning in predicting bus passengers based on simulated annealing, Appl. Soft Comput., 88 (2020), 106068. https://doi.org/10.1016/j.asoc.2020.106068 doi: 10.1016/j.asoc.2020.106068
[45]	Q. Yan, J. Chen, L. De Strycker, An outlier detection method based on Mahalanobis distance for source localization, Sensors, 18 (2018), 2186. https://doi.org/10.3390/s18072186 doi: 10.3390/s18072186
[46]	B. Mishra, T. B. Shahi, Deep learning-based framework for spatiotemporal data fusion: an instance of landsat 8 and sentinel 2 NDVI, J. Appl. Remote Sens., 15 (2021), 034520. https://doi.org/10.1117/1.JRS.15.034520 doi: 10.1117/1.JRS.15.034520

This article has been cited by:

Ahmed A. Gaber, Abdul-Majid Wazwaz, Dynamic wave solutions for (2+1)-dimensional DJKM equation in plasma physics, 2024, 9, 2473-6988, 6060, 10.3934/math.2024296

Reader Comments

Your name:*

Email:*
© 2023 the Author(s), licensee AIMS Press. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0)