Virtual target screening to rapidly identify potential protein targets of natural products in drug discovery

Yuri Pevzner; Daniel N. Santiago; Jacqueline L. von Salm; Rainer S. Metcalf; Kenyon G. Daniel; Laurent Calcul; H. Lee Woodcock; Bill J. Baker; Wayne C. Guida; Wesley H. Brooks; Yuri Pevzner; Daniel N. Santiago; Jacqueline L. von Salm; Rainer S. Metcalf; Kenyon G. Daniel; Laurent Calcul; H. Lee Woodcock; Bill J. Baker; Wayne C. Guida; Wesley H. Brooks

doi:10.3934/molsci.2014.2.81

AIMS Molecular Science

2014, Volume 1, Issue 2: 81-98. doi: 10.3934/molsci.2014.2.81

Previous Article Next Article

Research article Special Issues

Virtual target screening to rapidly identify potential protein targets of natural products in drug discovery

1.
Department of Chemistry, University of South Florida, Tampa, FL 33620, USA;
2.
Center for Molecular Diversity in Drug Design, Discovery and Delivery, University of South Florida, Tampa, FL 33620, USA;
3.
Virtual Screening and Molecular Modeling Core, H. Lee Moffitt Cancer Center & Research Institute, Tampa, FL 33612, USA;
4.
Department of Cell Biology, Microbiology, and Molecular Biology, University of South Florida, Tampa, FL 33620, USA;
5.
The Center for Drug Discovery and Innovation (CDDI), University of South Florida, Tampa, FL 33620, USA

Received: 17 March 2014 Accepted: 14 May 2014 Published: 21 May 2014

Inherent biological viability and diversity of natural products make them a potentially rich source for new therapeutics. However, identification of bioactive compounds with desired therapeutic effects and identification of their protein targets is a laborious, expensive process. Extracts from organism samples may show desired activity in phenotypic assays but specific bioactive compounds must be isolated through further separation methods and protein targets must be identified by more specific phenotypic and in vitro experimental assays. Still, questions remain as to whether all relevant protein targets for a compound have been identified. The desire is to understand breadth of purposing for the compound to maximize its use and intellectual property, and to avoid further development of compounds with insurmountable adverse effects. Previously we developed a Virtual Target Screening system that computationally screens one or more compounds against a collection of virtual protein structures. By scoring each compound-protein interaction, we can compare against averaged scores of synthetic drug-like compounds to determine if a particular protein would be a potential target of a compound of interest. Here we provide examples of natural products screened through our system as we assess advantages and shortcomings of our current system in regards to natural product drug discovery.

Keywords:

Citation: Yuri Pevzner, Daniel N. Santiago, Jacqueline L. von Salm, Rainer S. Metcalf, Kenyon G. Daniel, Laurent Calcul, H. Lee Woodcock, Bill J. Baker, Wayne C. Guida, Wesley H. Brooks. Virtual target screening to rapidly identify potential protein targets of natural products in drug discovery[J]. AIMS Molecular Science, 2014, 1(2): 81-98. doi: 10.3934/molsci.2014.2.81

Related Papers:

[1]	Yaxin Hou, Cao Wen, Yang Liu, Hong Li . A two-grid ADI finite element approximation for a nonlinear distributed-order fractional sub-diffusion equation. Networks and Heterogeneous Media, 2023, 18(2): 855-876. doi: 10.3934/nhm.2023037
[2]	Jinhu Zhao . Natural convection flow and heat transfer of generalized Maxwell fluid with distributed order time fractional derivatives embedded in the porous medium. Networks and Heterogeneous Media, 2024, 19(2): 753-770. doi: 10.3934/nhm.2024034
[3]	Ming-Kai Wang, Cheng Wang, Jun-Feng Yin . A second-order ADI method for pricing options under fractional regime-switching models. Networks and Heterogeneous Media, 2023, 18(2): 647-663. doi: 10.3934/nhm.2023028
[4]	Yin Yang, Aiguo Xiao . Dissipativity and contractivity of the second-order averaged L1 method for fractional Volterra functional differential equations. Networks and Heterogeneous Media, 2023, 18(2): 753-774. doi: 10.3934/nhm.2023032
[5]	Diandian Huang, Xin Huang, Tingting Qin, Yongtao Zhou . A transformed $ L1 $ Legendre-Galerkin spectral method for time fractional Fokker-Planck equations. Networks and Heterogeneous Media, 2023, 18(2): 799-812. doi: 10.3934/nhm.2023034
[6]	Panpan Xu, Yongbin Ge, Lin Zhang . High-order finite difference approximation of the Keller-Segel model with additional self- and cross-diffusion terms and a logistic source. Networks and Heterogeneous Media, 2023, 18(4): 1471-1492. doi: 10.3934/nhm.2023065
[7]	Jie Gu, Lijuan Nong, Qian Yi, An Chen . Two high-order compact difference schemes with temporal graded meshes for time-fractional Black-Scholes equation. Networks and Heterogeneous Media, 2023, 18(4): 1692-1712. doi: 10.3934/nhm.2023074
[8]	Wen Dong, Dongling Wang . Mittag-Leffler stability of numerical solutions to linear homogeneous time fractional parabolic equations. Networks and Heterogeneous Media, 2023, 18(3): 946-956. doi: 10.3934/nhm.2023041
[9]	Farman Ali Shah, Kamran, Dania Santina, Nabil Mlaiki, Salma Aljawi . Application of a hybrid pseudospectral method to a new two-dimensional multi-term mixed sub-diffusion and wave-diffusion equation of fractional order. Networks and Heterogeneous Media, 2024, 19(1): 44-85. doi: 10.3934/nhm.2024003
[10]	Zhe Pu, Maohua Ran, Hong Luo . Effective difference methods for solving the variable coefficient fourth-order fractional sub-diffusion equations. Networks and Heterogeneous Media, 2023, 18(1): 291-309. doi: 10.3934/nhm.2023011

Abstract

1. Introduction

In this paper, we consider the following two-dimensional multi-term time fractional diffusion equation:

$\begin{align} &{D}_{t}^{\mathit{\boldsymbol{\alpha}}} u(x,y,t)- \Delta u(x,y,t) = f(x,y,t) \quad \forall(x,y,t) \in Q: = \Omega \times(0, T], \end{align}$ $

(1.1a)

$\begin{align} &\left.u(x,y,t)\right|_{\partial \Omega} = 0 \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ {\text { for }} t \in(0, T], \end{align}$ $

(1.1b)

$\begin{align} &u(x,y,0) = u_{0}(x,y) \quad \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ {\text { for }} (x,y) \in \Omega, \end{align}$ $

(1.1c)

where $ f\in C(\bar{Q}) $, $ \Omega \subset \mathbb{R}^{2} $. In Eq(1.1a), $ {D}_{t}^{\mathit{\boldsymbol{\alpha}}} u $ denotes the multi-term fractional derivative, which is defined by

$\begin{equation} {D}_{t}^{\mathit{\boldsymbol{\alpha}}} u(x,y,t) = \sum\limits_{m = 1}^J b_m D_{t}^{\alpha_m} u(x,y,t), \end{equation}$ $

(1.2)

where $ J $ is a positive integer, $ b_m \geq 0 $, $ \mathit{\boldsymbol{\alpha}}: = (\alpha_1, \cdots, \alpha_J) $ with $ 0 < \alpha_J < \cdots < \alpha_1 < 1 $, and the fractional Caputo derivative $ D_{t}^{\alpha} u (0 < \alpha < 1) $ in Eq (1.2) is defined by

$\begin{equation*} D_{t}^{\alpha} u(x,y,t) = \frac{1}{\Gamma(1-\alpha)} \int_{0}^{t}(t-s)^{-\alpha} \frac{\partial u(x,y,s)}{\partial s} ds, \quad t > 0 . \end{equation*}$ $

For existence and uniqueness of the exact solution for problem (1.1), one can refer to ^[15,17]. The initial weak singularity of the solution for problem (1.1) is given in ^[9]. In recent years, many authors have investigated single-term time-fractional diffusion equations, see, e.g., ^[16,28,30]. At the same time, multi-term fractional equations ^[12,22,26] which have been successfully used in application of real life have attracted more and more attention. Among many types of fractional derivatives, some researchers have used Riemann-Liouville derivative ^[7,25], while others, including this article, choose to use Caputo derivative ^[13,27,30]. However, most of the previous work assumes that the exact solutions in temporal direction are smooth enough, whereas the solutions typically exhibit some weak singularities at initial time. Alternating direction implicit (ADI) method was first introduced in ^[18], the advantage of the ADI method is that it can reduce the computational cost by transforming a multi-dimensional problem into sets of 1D problems. Nowadays, there are many researchers using ADI method to solve various types of fractional derivative problems, such as ^{[2,6,14,19,20,21,23,31]}. Recently, Huang et al. ^[11] have presented an ADI scheme for 2D multi-term time-space fractional nonlinear diffusion-wave equations under reasonable solution regularity assumption. Cao and Chen ^[1] have used ADI difference method on uniform mesh to solve a 2D multi-term time-fractional subdiffusion equation with initial singularity. However, the global accuracy in time direction of ^[1] is low. It is only $ O(\tau^\alpha) $. So, our current work is on graded mesh to improve the temporal accuracy. More precisely, we investigate a fully discrete ADI method for solving the problem (1.1) with weakly singular solution, where the temporal discretization is based L1 approximation on graded mesh, and finite difference method is used for spatial discretization. We establish the stability and convergence of the fully discrete L1-ADI scheme, both $ L^2 $-norm and $ H^1 $-norm error estimates are obtained, and the final error bounds do not blow up when $ \alpha_1 \to 1^- $.

The rest of the paper is organized as follows. In Section 2, we construct a fully discrete L1-ADI scheme for problem (1.1). In Section 3, we establish the stability and convergence L1-ADI scheme in discrete $ L^2 $-norm. Then, the sharp $ H^{1} $-norm convergence of L1-ADI scheme is presented in Section 4. Some numerical experiments are given in Section 5. The final part is the conclusion.

Notation. Throughout the paper, we denote by $ C $ a generic positive constant, which may change its value at different occurrences, but is always independent of the mesh sizes. We call a constant $ C $ $ \alpha $-robust, if it doesn't blow up when $ \alpha_1 \to 1^- $.

2. Fully discrete L1-ADI scheme

In the whole paper, to simplify the analysis, let us choose $ \Omega = (0, L)\times(0, L) $, where $ L > 0 $ is constant. We use positive integers $ N_{1} $, $ N_{2} $ and $ M $ respectively to define the spatial and temporal partition parameters. Consider the graded temporal grid in $ [0, T]: t_{j} = T(j/M)^r $, $ j = 0, 1, \ldots, M $ and $ r\geq1 $. Denote time step $ \tau_n: = t_n-t_{n-1} $ for $ n = 1, \cdots, M $. Then, we use $ x_{m} = mh_{1} $ for $ m = 0, 1, \ldots, N_{1} $ and $ y_{n} = nh_{2} $ for $ n = 0, 1, \ldots, N_{2} $ to denote the spatial grids, where $ h_{1} = L/N_{1} $, $ h_{2} = L/N_{2} $. Set $ \Omega^{\ast}_{h} = \{(x_{m}, y_{n})\lvert0\leq m \leq N_{1}, 0\leq n \leq N_{2}\} $, $ \Omega_{h} = \Omega^{\ast}_{h}\cap \Omega $, and $ \partial\Omega_{h} = \Omega^{\ast}_{h}\cap \partial\Omega $.

We approximate the Caputo fractional derivative $ D_{t}^{\alpha_{m}} u\left(\cdot, \cdot, t_{n}\right) $ with $ 0 < \alpha_{m} < 1 $ and $ 1 \leq n \leq M $ on graded mesh by using well-known L1 scheme as follows:

$\begin{equation} \delta_{t}^{\alpha_{m}} u^{n}: = z_{n, 1}^{(\alpha_{m})} u^{n}-z_{n, n}^{(\alpha_{m})} u^{0}-\sum\limits_{i = 1}^{n-1}\left(z_{n, i}^{(\alpha_{m})}-z_{n, i+1}^{(\alpha_{m})}\right) u^{n-i}, \end{equation}$ $

(2.1)

where

$\begin{equation} z_{n,k}^{(\alpha_{m})} = \frac{1}{\Gamma\left(2-\alpha_{m}\right)}\frac{\left(t_{n}-t_{n-k}\right)^{1-\alpha_{m}}-\left(t_{n}-t_{n-k+1}\right)^{1-\alpha_{m}}}{\tau_{n-k+1}}\quad k = 1,2,\ldots,n. \end{equation}$ $

(2.2)

Then, we denote

$\begin{equation} z_{n,i}: = \sum\limits_{m = 1}^{J} b_m z_{n,i}^{\left(\alpha_{m}\right)}. \end{equation}$ $

(2.3)

Thus one can approximate the multi-term fractional derivative $ D_t^{\mathit{\boldsymbol{\alpha}}} u $ in Eq (1.2) by

$\begin{equation} \delta_{t}^{\mathit{\boldsymbol{\alpha}}} u^{n}: = z_{n,1} u^{n}-z_{n,n}u^{0}-\sum\limits_{k = 1}^{n-1}\left(z_{n,k}-z_{n, k+1}\right) u^{n-k}. \end{equation}$ $

(2.4)

Given a mesh function$ \{v^j\}_{j = 0}^M $, and for $ 0\leq n\leq M $, we define

$\begin{align*} &\delta_{x}v_{i-\frac{1}{2},j}^n = \frac{v_{i,j}^n-v_{i-1,j}^n}{h_{1}} \ \ \; \; \text{ for } 1\leq i\leq N_1, 0\leq j\leq N_2,\\ &\delta_{x}^{2}v^n_{i,j} = \frac{\delta_{x}v^n_{i+\frac{1}{2},j}-\delta_{x}v^n_{i-\frac{1}{2},j}}{h_{1}} \ \ \; \; \text{ for } 1\leq i\leq N_1-1, 0\leq j\leq N_2, \\ &\delta_{y}\delta_{x}v^n_{i-\frac{1}{2},j-\frac{1}{2}} = \frac{\delta_{x}v^n_{i-\frac{1}{2},j}-\delta_{x}v^n_{i-\frac{1}{2},j-1}}{h_{2}} \ \ \; \; \text{ for } 1\leq i\leq N_1, 1\leq j\leq N_2. \end{align*}$ $

The notations $ \delta_{y}v^n_{i, j-\frac{1}{2}} $, $ \delta_{y}^{2}v^n_{i, j} $ and $ \delta_{x}\delta_{y}v^n_{i-\frac{1}{2}, j-\frac{1}{2}} $ can be defined similarly. We define the discrete Laplace operator $ \Delta_{h}: = \delta_{x}^{2}+\delta_{y}^{2} $ which is a second-order approximation of $ \Delta $. Let $ u_{i, j}^{n} $ be the numerical approximation of the exact solution $ u(x_{i}, y_{j}, t_{n}) $, so the discrete problem of (1.1) is as follows

$\begin{align} &\delta_{t}^{\mathit{\boldsymbol{\alpha}}} u_{i,j}^{n}- \Delta_{h} u_{i,j}^{n} = f_{i,j}^{n}\quad \forall(x_{i},y_{j},t_{n}) \in Q: = \Omega \times(0, T], \end{align}$ $

(2.5a)

$\begin{align} &u_{i,j}^{n}|_{\partial \Omega} = 0 \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ {\text { for }} t \in(0, T], \end{align}$ $

(2.5b)

$\begin{align} &u_{i,j}^{0} = u_{0}(x_{i},y_{j}) \quad \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ {\text { for }} (x_{i},y_{j}) \in \Omega. \end{align}$ $

(2.5c)

To solve 2D problem, we want to solve 1D problem at first, after that we solve another 1D problem. If a small term $ \gamma_{n}^{2}\delta_{x}^{2}\delta_{y}^{2}\delta_{t}^{\mathit{\boldsymbol{\alpha}}} u_{i, j}^{n} $ for $ n = 1, \ldots, M $, where $ \gamma_{n} = z_{n, 1}^{-1} $, is added to the left side of Eq (2.5a), one gets

$\begin{equation} (1+\gamma_{n}^{2}\delta_{x}^{2}\delta_{y}^{2})\delta_{t}^{\mathit{\boldsymbol{\alpha}}} u_{i,j}^{n}-\delta_{x}^{2}u_{i,j}^{n}-\delta_{y}^{2}u_{i,j}^{n} = f_{i,j}^{n}\quad \forall(x_{i},y_{j},t_{n}) \in Q: = \Omega \times(0, T]. \\ \end{equation}$ $

(2.6)

Thus the purpose is achieved as Eq (2.6) can be rewritten by

$\begin{equation*} (1-\gamma_{n}\delta_{x}^{2})(1-\gamma_{n}\delta_{y}^{2}) u_{i,j}^{n} = \gamma_{n}[(1+\gamma_{n}^{2}\delta_{x}^{2}\delta_{y}^{2})(z_{n,n}u_{i,j}^{0}+\sum\limits_{k = 1}^{n-1}(z_{n,k}-z_{n,k+1})u_{i,j}^{n-k})+f_{i,j}^{n}]. \end{equation*}$ $

We set $ u_{i, j}^{*} = (1-\gamma_{n}\delta_{y}^{2})u_{i, j} $. Then, the first 1D problem we need to solve is, for $ 1\leq j\leq N_2-1 $,

$\begin{equation} \left\{ \begin{aligned} &(1-\gamma_{n}\delta_{x}^{2})u_{i,j}^{*} = \gamma_{n}[(1+\gamma_{n}^{2}\delta_{x}^{2}\delta_{y}^{2})(z_{n,n}u_{i,j}^{0}+\sum\limits_{k = 1}^{n-1}(z_{n,k}-z_{n,k+1})u_{i,j}^{n-k})+f_{i,j}^{n}] \ \ \ \ 1 \leq i \leq N_{1}-1,\\ &u_{0,j}^{*} = (1-\gamma_{n}\delta_{y}^{2})u_{0,j}^{n}, u_{N_{1},j}^{*} = (1-\gamma_{n}\delta_{y}^{2})u_{N_{1},j}^{n}, \end{aligned} \right. \end{equation}$ $

(2.7)

and the second 1D problem we need to solve is, for $ 1\leq i\leq N_1-1 $,

$\begin{equation} \left\{ \begin{aligned} &(1-\gamma_{n}\delta_{y}^{2})u_{i,j}^{n} = u_{i,j}^{*} \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ 1 \leq j\leq N_{2}-1,\\ &u_{i,0}^{n} = 0, u_{i,N_{2}}^{n} = 0. \end{aligned} \right. \end{equation}$ $

(2.8)

Thus, we have the following fully discrete ADI scheme for the problem (1.1):

$\begin{align} &(1+\gamma_{n}^{2}\delta_{x}^{2}\delta_{y}^{2})\delta_{t}^{\mathit{\boldsymbol{\alpha}}} u_{i,j}^{n}-\delta_{x}^{2}u_{i,j}^{n}-\delta_{y}^{2}u_{i,j}^{n} = f_{i,j}^{n}\quad \forall(x_{i},y_{j},t_{n}) \in Q: = \Omega \times(0, T], \end{align}$ $

(2.9a)

$\begin{align} &u_{i,j}^{n}|_{\partial \Omega} = 0 \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ {\text { for }} t \in(0, T], \end{align}$ $

(2.9b)

$\begin{align} &u_{i,j}^{0} = u_{0}(x_{i},y_{j}) \quad \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ {\text { for }} (x_{i},y_{j}) \in \Omega. \end{align}$ $

(2.9c)

3. Analysis of stability and convergence

We define the convolution multipliers $ \sigma_{n, j} $, which is positive for $ n = 1, 2, \ldots, M $ and $ j = 1, 2, \ldots, n-1 $ by

$\begin{equation} \sigma_{n, n} = 1, \quad \sigma_{n, j} = \sum\limits_{k = 1}^{n-j} \frac{1}{z_{n-k,1}}\left(z_{n,k}-z_{n,k+1}\right) \sigma_{n-k, j} > 0 . \end{equation}$ $

(3.1)

We have the following two lemmas on the properties of the convolution multipliers $ \sigma_{n, j} $.

Lemma 1. ([10, Corollary 1]) One has

$ z_{n,1}^{-1}\sum\limits_{j = 1}^n\sigma_{n,j}\leq \sum\limits_{m = 1}^J\frac{t_n^{\alpha_m}}{b_m{\Gamma(1+\alpha_m)}}. $

Lemma 2. ([10, Corollary 2]) Set $ l_M = 1/\ln M $. Assume that $ M\ge 3 $ so $ 0 < l_M < 1 $. Then

$ z_{n,1}^{-1}\sum\limits_{j = 1}^n \left(\sum _{m = 1}^Jb_mt_j^{-\alpha_m}\right)\sigma _{n,j}\leq \frac{Je^r\max\limits_{1\leq m\leq J} \Gamma(1+l_M-\alpha_m)}{\Gamma(1+l_M)}. $

To investigate the stability and convergence of the fully discrete L1-ADI scheme (2.9), we need the following fractional discrete Gronwall inequality.

Lemma 3. [8, Lemma 5.3] Suppose that the sequences $ \left\{\varepsilon^n\right\}_{n = 1}^{\infty}, \left\{\eta^n\right\}_{n = 1}^{\infty} $ are nonnegative and assume the grid function $ \left\{V^n: n = 0, 1, \ldots, M\right\} $ satisfies $ V_0 \geq 0 $ and

$ \left(\delta_t^{\boldsymbol{\alpha}} V^n\right) V^n \leq \varepsilon^n V^n+\left(\eta^n\right)^2 \quad for \quad n = 1,2, \ldots, M. $

Then,

$\begin{equation*} V^n \leq V^0+z_{n,1}^{-1} \sum\limits_{j = 1}^n \sigma_{n,j}\left(\varepsilon^n+\eta^n\right)+\max\limits _{1 \leq j \leq n}\left\{\eta^j\right\} \quad for \quad n = 1,2, \ldots, M. \end{equation*}$ $

Set

$ \mathcal{V}_h = \big\{ v\lvert v = \{v_{i,j}\lvert(x_i,y_j)\in \Omega_h \} \quad\text{and} \quad v_{i,j} = 0 \quad\text{if}\quad (x_i,y_j)\in \partial\Omega_h \big\}. $

For any mesh functions $ u, v \in \mathcal{V}_h $, define the discrete $ L^2 $ inner product $ (u, v) = h_1h_2\sum_{i = 1}^{N_1-1}\sum_{j = 1}^{N_2-1}u_{i, j}v_{i, j} $, and the norm $ \Vert v \Vert = \sqrt{(v, v)} $. We also define a new inner product $ (u, v)_{\gamma_{n}}: = (u, v)+ $ $ \gamma_{n}^2(u, v)_{xy} $, where

$ (u, v)_{x y} = h_1 h_2 \sum\limits_{i = 1}^{N_1} \sum\limits_{j = 1}^{N_2}\left(\delta_x \delta_y u_{i-\frac{1}{2}, j-\frac{1}{2}}\right) \delta_x \delta_y v_{i-\frac{1}{2}, j-\frac{1}{2}}, $

and set $ \|u\|_{\gamma_{n}} = \sqrt{(u, u)_{\gamma_{n}}} $.

Lemma 4. [1, Lemma 3] The inner product $ (u, v)_{\gamma_{n}} $ satisfies

$\begin{equation*} (u, v)_{\gamma_{n}} \leq\|u\|_{\gamma_{n}}\|v\|_{\gamma_{n}} . \end{equation*}$ $

Lemma 5. For $ n = 1, 2, \ldots, M $, the solution of ADI scheme (2.9) satisfies

$\begin{equation*} \left\|u^n\right\|_{\gamma_{n}} \leq\left\|u^0\right\|_{\gamma_{n}}+z_{n,1}^{-1} \sum\limits_{j = 1}^n \sigma_{n,j}\left\|f^j\right\|. \end{equation*}$ $

Proof. On both sides of Eq (2.9a), we take the discrete inner product with $ u^n $. Then, one has

$\begin{equation} \begin{aligned} &\left(\left(1+\gamma_{n}^2 \delta_x^2 \delta_y^2\right)\left(z_{n,1} u^n-\sum\limits_{k = 1}^{n-1}\left(z_{n,k}-z_{n,k+1}\right) u^{n-k}-z_{n,n} u^0\right), u^n\right) \\ = &\left(\delta_x^2 u^n, u^n\right)+\left(\delta_y^2 u^n, u^n\right)+\left(f^n, u^n\right). \end{aligned} \end{equation}$ $

(3.2)

Then, by linear property of the inner product and discrete Green formula, the Eq (3.2) becomes

$\begin{equation*} \begin{aligned} z_{n,1}\left\|u^n\right\|_{\gamma_{n}}^2-\sum\limits_{k = 1}^{n-1}\left(z_{n,k}-z_{n,k+1}\right) (u^{n-k},u^n)_{\gamma_{n}}-z_{n,n}( u^0, u^n)_{\gamma_{n}} = \left(f^n, u^n\right)-\|\delta_x u^n\|^2-\|\delta_y u^n\|^2. \end{aligned} \end{equation*}$ $

Using Lemma 4, one has

$\begin{equation*} z_{n,1}\left\|u^n\right\|_{\gamma_{n}}^2 \leq \sum\limits_{k = 1}^{n-1}\left(z_{n,k}-z_{n,k+1}\right)\left\|u^{n-k}\right\|_{\gamma_{n}} \left\|u^n\right\|_{\gamma_{n}}+z_{n,n}\left\|u^0\right\|_{\gamma_{n}}\left\|u^n\right\|_{\gamma_{n}}+\left\|f^n\right\|\left\|u^n\right\|_{\gamma_{n}}, \end{equation*}$ $

which is equivalent to

$\begin{equation} \left(\delta_t^\alpha\left\|u^n\right\|_{\gamma_{n}}\right)\left\|u^n\right\|_{\gamma_{n}} \leq \left\|f^n\right\|\left\|u^n\right\|_{\gamma_{n}}. \end{equation}$ $

(3.3)

Then, setting all $ \eta^n = 0 $ in Lemma 3, and applying it to Eq (3.3), the result is proved. □

From [3, Lemmas 2.2 and 2.3], one can easily get that

Lemma 6. Set $ \sigma \in(0, 1) $. Suppose that $ \left|u^{(k)}(t)\right| \leq C\left(1+t^{\sigma-k}\right) $ where $ k = 0, 1, 2 $. Then, for $ 1\leq n\leq M $, one has

$\begin{equation*} \begin{aligned} \left|\delta_{t}^{\mathit{\boldsymbol{\alpha}}} u\left(t_{n}\right)- D_{t}^{\mathit{\boldsymbol{\alpha}}} u\left(t_{n}\right)\right| \leq C \sum\limits_{m = 1}^{J} b_m t_{n}^{-\alpha_{m}} M^{-\min \left\{r\sigma, 2-\alpha_{m}\right\}}. \end{aligned} \end{equation*}$ $

Now we come to the convergence of the fully discrete ADI scheme (2.9).

Theorem 1. Suppose that {$ \left|u^{(l)}(t)\right| \leq C\left(1+t^{\sigma-l}\right) $} for $ l = 0, 1, 2 $ with $ \sigma \in(0, 1) $. Then the computed solution errors $ e_{i, j}^n: = u\left(x_i, y_j, t_n\right)-u_{i, j}^n $ satisfy:

$\begin{equation*} \left\|e^n\right\| \leq C \left(h_1^2+h_2^2+M^{-\min \left\{ 2\alpha_{1}, r\sigma, 2-\alpha_1\right\}}\right),\; n = 1,2, \ldots, M, \end{equation*}$ $

where C is $ \alpha $-robust.

Proof. From Eqs (2.9a) and (1.1a), we get the following error equation:

$\begin{equation} \left(1+\gamma_{n}^2 \delta_x^2 \delta_y^2\right) \delta_t^\alpha e_{i, j}^n-\delta_x^2 e_{i, j}^n-\delta_y^2e_{i, j}^n = \phi_{i, j}^n\ \ {\rm{for }}\;n = 1,2, \ldots, M, \end{equation}$ $

(3.4)

where $ \phi_{i, j}^n $ is the truncation error. By the well-known simple bound $ |\Delta u-\Delta_h u|\leq C(h_1^{2}+h_2^2) $, $ |\gamma_{n}^2| \leq C M^{-2\alpha_1} $, and Lemma $ 6 $, the truncation error $ \phi_{i, j}^n $ satisfies

$\begin{aligned} \left|\phi_{i, j}^n\right| \leq C\left(h_1^{2}+h_{2}^{2}+M^{-2\alpha_1}+\sum\limits_{m = 1}^{J} b_m t_{n}^{-\alpha_{m}} M^{-\min \left\{r \sigma, 2-\alpha_{m}\right)}\right). \end{aligned}$ $

From Lemma 5, and noting that $ e^0 = 0 $, we have

$\begin{aligned} \left\|e^n\right\|_{\gamma_{n}} & \leq z_{n,1}^{-1} \sum\limits_{j = 1}^n \sigma_{n,j}\left\|\phi^j\right\| \\ & \leq Cz_{n,1}^{-1} \sum\limits_{j = 1}^n \sigma_{n,j}\left(h_1^{2}+h_{2}^{2}+M^{-2\alpha_1}+\sum\limits_{m = 1}^{J} b_m t_{n}^{-\alpha_{m}} M^{-\min \left\{r \sigma, 2-\alpha_{m}\right)}\right). \end{aligned}$ $

Then, using Lemmas 1 and 2, we can get

$\begin{aligned} \left\|e^n\right\|_{\gamma_{n}} \leq C \left(h_1^2+h_2^2+M^{-\min \left\{ 2\alpha_{1}, r\sigma, 2-\alpha_1\right\}}\right), \end{aligned}$ $

where $ C $ is $ \alpha $-robust. Finally, according to the definition of the norm $ \left\|\cdot\right\|_{\gamma_{n}} $, the proof is completed. □

4. $ H^1 $-norm convergence of L1-ADI scheme

In order to prove the stability and convergence of the fully discrete ADI scheme (2.9) in $ H^1 $-norm sense in this section, we first define some norms. For any grid functions $ U, V \in \mathcal{V}_h $, define

$\begin{aligned} & (U, V)_{x y^2} = h_1 h_2 \sum\limits_{i = 1}^{M_1-1} \sum\limits_{j = 1}^{M_2}\left(\delta_x \delta_y^2 U_{i, j-\frac{1}{2}}\right) \delta_x \delta_y^2 V_{i, j-\frac{1}{2}}, \\ & (U, V)_{x^2 y} = h_1 h_2 \sum\limits_{i = 1}^{M_1} \sum\limits_{j = 1}^{M_2-1}\left(\delta_x^2 \delta_y U_{i-\frac{1}{2}, j}\right) \delta_x^2 \delta_y V_{i-\frac{1}{2}, j}, \\ & (U, V)_{\Delta_h} = h_1 h_2 \sum\limits_{i = 1}^{M_1} \sum\limits_{j = 1}^{M_2}\left(\Delta_h U_{i, j}\right) \Delta_h V_{i, j} . \end{aligned}$ $

The corresponding seminorms are

$\begin{aligned} \left\|\delta_x \delta_y^2 U\right\| & = \sqrt{(U, U)_{x y^2}}, \\ \left\|\delta_x^2 \delta_y U\right\| & = \sqrt{(U, U)_{x^2 y}}, \\ \left\|\Delta_h U\right\| & = \sqrt{(U, U)_{\Delta_h}}. \end{aligned}$ $

For any function $ V \in \mathcal{V}_h $, define

$\begin{aligned} \left\|\nabla_h V^n\right\| & = \sqrt{\left\|\delta_x V^n\right\|^2+\left\|\delta_y V^n\right\|^2}, \\ \|V^n\|_{H^1} & = \sqrt{\|V^n\|^2+\left\|\nabla_h V^n\right\|^2}, \\ \|V^n\|_A & = \sqrt{\left\|\nabla_h V\right\|^2+\gamma_n^2\left(\left\|\delta_x \delta_y^2 V^n\right\|^2+\left\|\delta_x^2 \delta_y V^n\right\|^2\right)}. \end{aligned}$ $

According to the definition of $ \left\|\nabla_h V^n\right\| $ and $ \|V^n\|_A $, we can see that $ \left\|\nabla_h V^n\right\| \leq \|V^n\|_A $. From [31, Lemma 2.2], one has $ \|V^n\| \leq C\left\|\nabla_h V^n\right\| $ for all functions $ V \in \mathcal{V}_h $, then, we have $ \|V^n\|_{H^1} \leq C\left\|\nabla_h V^n\right\| $, hence, $ \|V^n\|_{H^1} \leq C\|V^n\|_A $.

Lemma 7. [29, Lemma 3.2] For any grid functions $ U, V \in \mathcal{V}_h $, one has

$\begin{equation*} \left(-\left(U^n+\gamma_{n}^2 \delta_x^2 \delta_y^2 U^n\right), \Delta_h V^n\right) \leq\|U^n\|_A\|V^n\|_A, \end{equation*}$ $

where the equality holds when $ U = V $.

Next, we denote $ R_t u_{i, j}^n = \left(1+\gamma_{n}^2 \delta_x^2 \delta_y^2\right) \delta_t^{\mathit{\boldsymbol{\alpha}}} u\left(x_i, y_j, t_n\right)-D_t^{\mathit{\boldsymbol{\alpha}}} u\left(x_i, y_j, t_n\right) $, $ R_s u_{i, j}^n = \Delta u\left(x_i, y_j, t_n\right)-\Delta_h u\left(x_i, y_j, t_n\right) $. Then the error equation (3.4) can be written as

$\begin{equation} \left(1+{\gamma_{n}}^2 \delta_x^2 \delta_y^2\right) \delta_t^{\boldsymbol{\alpha}} e_{i, j}^n-\Delta_h e_{i, j}^n = R_t u_{i, j}^n+R_s u_{i, j}^n. \end{equation}$ $

(4.1)

Besides, we have

$\begin{equation} \left(R_t u^n,-\Delta_h e^n\right) \leq\left\|\nabla_h R_t u^n\right\|\left\|\nabla_h e^n\right\| \leq\left\|\nabla_h R_t u^n\right\|\left\|e^n\right\|_A \end{equation}$ $

(4.2)

and

$\begin{equation} \left(R_s u^n,-\Delta_h e^n\right)\leq\left\| R_s u^n\right\|\left\|\Delta_h e^n\right\| \leq\left\|\Delta_h e^n\right\|^2+\frac{1}{4}\left\|R_s u^n\right\|^2. \end{equation}$ $

(4.3)

Theorem 2. Suppose that $ \left|u^{(l)}(t) \right|\leq C\left(1+t^{\sigma-l}\right) $ for $ l = 0, 1, 2 $ with $ \sigma \in(0, 1) $. Then, the computed solution error $ e_{i, j}^n: = u\left(x_i, y_j, t_n\right)-u_{i, j}^n $ satisfy:

$\begin{equation*} \left\|e^n\right\|_{H^1} \leq C \left(h_1^2+h_2^2+M^{-\min \left\{ 2\alpha_{1}, r\sigma, 2-\alpha_1\right\}}\right) \quad for \ n = 1,2, \ldots, M, \end{equation*}$ $

where C is $ \alpha $-robust.

Proof. Taking discrete $ L^2 $ inner product with $ -\Delta_h e^n $ on both sides of Eq (4.1) and noting that $ e^0 = 0 $, we get

$\begin{aligned} & z_{n,1}\left(e^n+\gamma_{n}^2 \delta_x^2 \delta_y^2 e^n,-\Delta_h e^n\right)+\left\|\Delta_h e^n\right\|^2 \\ = & \sum\limits_{k = 1}^{n-1}\left(z_{n,k}-z_{n,k+1}\right)\left(e^{n-k}+\gamma_{n}^2 \delta_x^2 \delta_y^2 e^{n-k},-\Delta_h e^n\right) +\left(R_t u^n+R_s u^n,-\Delta_h e^n\right). \end{aligned}$ $

Using Lemma 7, one has

$\begin{equation} z_{n,1}\left\|e^n\right\|_A^2+\left\|\Delta_h e^n\right\|^2 \leq \sum\limits_{k = 1}^{n-1}\left(z_{n,k}-z_{n,k+1}\right)\left\|e^{n-k}\right\|_A\left\|e^n\right\|_A+\left(R_t u^n+R_s u^n,-\Delta_h e^n\right). \end{equation}$ $

(4.4)

From Eqs (4.2) and (4.3), we obtain

$\begin{aligned} z_{n,1}\left\|e^n\right\|_A^2 \leq \sum\limits_{k = 1}^{n-1}\left(z_{n,k}-z_{n,k+1}\right)\left\|e^{n-k}\right\|_A\left\|e^n\right\|_A+\left\|\nabla_h R_t u^n\right\|\left\|e^n\right\|_A+\frac{1}{4}\left\|R_s u^n\right\|^2, \end{aligned}$ $

that is

$\begin{equation} \left(\delta_t^{\boldsymbol{\alpha}}\left\|e^n\right\|_A\right)\left\|e^n\right\|_A \leq\left\|\nabla_h R_t u^n\right\|\left\|e^n\right\|_A+\frac{1}{4}\left\|R_s u^n\right\|^2. \end{equation}$ $

(4.5)

Thus, by Lemma 3, from Eq (4.5) we have

$\begin{align} \|e^n\|_A \leq z_{n,1}^{-1} \sum\limits_{j = 1}^n \sigma_{n,j}\left(\left\|\nabla_h R_t u^j\right\|+\frac{1}{2}\left\|R_s u^j\right\|\right)+\max _{1 \leq j \leq n}\left\{\frac{1}{2}\left\|R_s u^j\right\|\right\}. \end{align}$ $

(4.6)

Then, using Lemmas 1 and 2, we come to the conclusion

$\begin{equation} \left\|e^n\right\|_A \leq C \left(h_1^2+h_2^2+M^{-\min \left\{ 2\alpha_{1}, r\sigma, 2-\alpha_1\right\}}\right) \quad for \ n = 1,2, \ldots, M \end{equation}$ $

(4.7)

by noting that

$ \left\|R_s u^j\right\|\leq C(h_1^2+h_2^2) $

and

$ \left\|\nabla_h R_t u^j\right\| \leq C\left( M^{-2\alpha_1}+\sum\limits_{m = 1}^{J} b_m t_{j}^{-\alpha_{m}} M^{-\min \left\{r \sigma, 2-\alpha_{m}\right)}\right). $

Finally, the results follow from $ \left\|e^n\right\|_{H^1} \leq C\left\|e^n\right\|_A $ for $ n = 1, 2, \ldots, M $. The proof is completed. □

5. Numerical experiments

In this section, some numerical examples are reported to support our theoretical analysis. In this section, we present some 2D numerical examples to illustrate the result of our error analysis and convergence order on the temporal graded mesh. We define $ \max\limits_{1 \leq n \leq M} \left\|u\left(t_n\right)-u^n\right\| $ as temporal global $ L^2 $-norm error and global $ H^1 $-norm error is defined similarly.

Example 1. In Eq (1.1) take $ \Omega = [0, \pi]\times[0, \pi] $, $ T = 1 $, $ J = 2 $, $ b_1 = b_2 = 1 $. Choose $ u(x, y, t) = t^{\alpha_{1}}\sin x\sin y $ as an exact solution. The force term $ f(x, y, t) = b_1\Gamma(1+\alpha_{1})+b_2\Gamma(1+\alpha_{1})\frac{t^{\alpha_{1}-\alpha_{2}}}{\Gamma(1-\alpha_{2}+\alpha_{1})}\sin x\sin y+2t^{\alpha_{1}}\sin x\sin y $.

For this example, the regularity parameter $ \sigma $ in Lemma 6 is $ \sigma = \alpha_{1} $. Thus, both the temporal convergence orders in Theorems 1 and 2 are $ O(M^{-\min \left\{ 2\alpha_{1}, r\alpha_1, 2-\alpha_1\right\}}) $. We take $ N_1 = N_2 = 1000 $ to eliminate the contamination of spatial errors. Tables 1–3 present $ L^2 $-norm errors and orders of convergence for Example 1 with different $ \alpha_1 $, $ \alpha_{2} $ and $ r $. And Tables 4, 6 and 8 present $ H^1 $-norm errors and convergence orders. From these tables, we find that the temporal convergence orders are consistent with our theoretical results. As in ^[1], we also present the local $ H^1 $-norm errors and orders of convergence at $ t = 1 $ for Example 1 in Tables 5, 7 and 9, which have better convergence rates than the global errors. We next test the spatial accuracy of the fully discrete ADI scheme (2.9). Take $ M = 1000 $ such that the spatial errors are dominant. We using $ \alpha_{1} = 0.6 $, $ \alpha_{2} = 0.4 $ and $ r = 2 $ as an example and the results are presented in Figure 1. The results show that the spatial convergence orders are second order in $ L^2 $ and $ H^1 $-norm sense as is indicated in Theorems 1 and 2.

Table 1. Global $ L^2 $-norm errors and orders of convergence for Example 1 with $ \alpha_1 = 0.4 $, $ \alpha_{2} = 0.2 $.

	$ M=64 $	$ M=128 $	$ M=256 $	$ M=512 $	$ M=1024 $	Theoretical order
$ r=1 $	4.5097e-02	3.5324e-02	2.7702e-02	2.1696e-02	1.6945e-02
$ r=1 $		0.3524	0.3507	0.3526	0.3565	0.4
$ r=2 $	1.6555e-02	1.0374e-02	6.4456e-03	3.9727e-03	2.4305e-03
$ r=2 $		0.6743	0.6866	0.6982	0.7089	0.8
$ r=\frac{1}{\alpha_1} $	1.8950e-02	1.1927e-02	7.4361e-03	4.5963e-03	2.8189e-03
$ r=\frac{1}{\alpha_1} $		0.6679	0.6816	0.6941	0.7053	0.8
$ r=\frac{2-\alpha_{1}}{\alpha_{1}} $	2.5296e-02	1.6072e-02	1.0095e-02	6.2790e-03	3.8718e-03
$ r=\frac{2-\alpha_{1}}{\alpha_{1}} $		0.6544	0.6708	0.6851	0.6975	0.8

| Show Table

DownLoad: CSV

Table 2. Global $ L^2 $-norm errors and orders of convergence for Example 1 with $ \alpha_1 = 0.6 $, $ \alpha_{2} = 0.4 $.

	$ M=64 $	$ M=128 $	$ M=256 $	$ M=512 $	$ M=1024 $	Theoretical order
$ r=1 $	2.1888e-02	1.4954e-02	1.0156e-02	6.8591e-03	4.6113e-03
$ r=1 $		0.5495	0.5583	0.5662	0.5728	0.6
$ r=2 $	4.3121e-03	2.0043e-03	9.2711e-04	4.2681e-04	1.9546e-04
$ r=2 $		1.1053	1.1123	1.1192	1.1267	1.2
$ r=\frac{1}{\alpha_1} $	4.6879e-03	2.4428e-03	1.2576e-03	6.4229e-04	3.2626e-04
$ r=\frac{1}{\alpha_1} $		0.9404	0.9578	0.9694	0.9772	1.0
$ r=\frac{2-\alpha_{1}}{\alpha_{1}} $	4.9526e-03	2.3114e-03	1.0727e-03	4.9519e-04	2.2734e-04
$ r=\frac{2-\alpha_{1}}{\alpha_{1}} $		1.0994	1.1076	1.1152	1.1231	1.2

| Show Table

DownLoad: CSV

Table 3. Global $ L^2 $-norm errors and orders of convergence for Example 1 with $ \alpha_1 = 0.8 $, $ \alpha_{2} = 0.6 $.

	$ M=64 $	$ M=128 $	$ M=256 $	$ M=512 $	$ M=1024 $	Theoretical order
$ r=1 $	7.1706e-03	4.2900e-03	2.5403e-03	1.4930e-03	8.7296e-04
$ r=1 $		0.7411	0.7560	0.7668	0.7743	0.8
$ r=2 $	1.9211e-03	7.5633e-04	2.9814e-04	1.1785e-04	4.6867e-05
$ r=2 $		1.3448	1.3430	1.3391	1.3302	1.2
$ r=\frac{1}{\alpha_1} $	3.7340e-03	1.9655e-03	1.0233e-03	5.2892e-04	2.7174e-04
$ r=\frac{1}{\alpha_1} $		0.9258	0.9417	0.9521	0.9608	1.0
$ r=\frac{2-\alpha_{1}}{\alpha_{1}} $	2.2479e-03	1.0660e-03	4.9903e-04	2.3116e-04	1.0617e-04
$ r=\frac{2-\alpha_{1}}{\alpha_{1}} $		1.0764	1.0950	1.1102	1.1225	1.2

| Show Table

DownLoad: CSV

Table 4. Global $ H^1 $-norm errors and orders of convergence for Example 1 with $ \alpha_1 = 0.4 $, $ \alpha_{2} = 0.2 $.

	$ M=64 $	$ M=128 $	$ M=256 $	$ M=512 $	$ M=1024 $	Theoretical order
$ r=1 $	4.5098e-02	3.5325e-02	2.7702e-02	2.1696e-02	1.6945e-02
$ r=1 $		0.3524	0.3507	0.3526	0.3565	0.4
$ r=2 $	1.6555e-02	1.0374e-02	6.4457e-03	3.9728e-03	2.4305e-03
$ r=2 $		0.6743	0.6866	0.6982	0.7089	0.8
$ r=\frac{1}{\alpha_1} $	1.8950e-02	1.1927e-02	7.4362e-03	4.5963e-03	2.8189e-03
$ r=\frac{1}{\alpha_1} $		0.6679	0.6816	0.6941	0.7053	0.8
$ r=\frac{2-\alpha_{1}}{\alpha_{1}} $	2.5296e-02	1.6072e-02	1.0095e-02	6.2790e-03	3.8718e-03
$ r=\frac{2-\alpha_{1}}{\alpha_{1}} $		0.6544	0.6708	0.6851	0.6975	0.8

| Show Table

DownLoad: CSV

Table 5. Local $ H^1 $-norm errors and orders of convergence at $ t = 1 $ for Example 1 with $ \alpha_1 = 0.4 $, $ \alpha_{2} = 0.2 $.

	$ M=64 $	$ M=128 $	$ M=256 $	$ M=512 $	$ M=1024 $
$ r=1 $	1.1675e-02	7.1582e-03	4.3712e-03	2.6568e-03	1.6067e-03
$ r=1 $		0.7057	0.7116	0.7184	0.7255
$ r=2 $	1.6555e-02	1.0374e-02	6.4457e-03	3.9728e-03	2.4305e-03
$ r=2 $		0.6743	0.6866	0.6982	0.7089
$ r=\frac{1}{\alpha_1} $	1.8950e-02	1.1927e-02	7.4362e-03	4.5963e-03	2.8189e-03
$ r=\frac{1}{\alpha_1} $		0.6679	0.6816	0.6941	0.7053
$ r=\frac{2-\alpha_{1}}{\alpha_{1}} $	2.5296e-02	1.6072e-02	1.0095e-02	6.2790e-03	3.8718e-03
$ r=\frac{2-\alpha_{1}}{\alpha_{1}} $		0.6544	0.6708	0.6851	0.6975

| Show Table

DownLoad: CSV

Table 6. Global $ H^1 $-norm errors and orders of convergence for Example 1 with $ \alpha_1 = 0.6 $, $ \alpha_{2} = 0.4 $.

	$ M=64 $	$ M=128 $	$ M=256 $	$ M=512 $	$ M=1024 $	Theoretical order
$ r=1 $	2.1888e-02	1.4955e-02	1.0156e-02	6.8591e-03	4.6113e-03
$ r=1 $		0.5495	0.5583	0.5662	0.5728	0.6
$ r=2 $	4.3121e-03	2.0043e-03	9.2712e-04	4.2681e-04	1.9546e-04
$ r=2 $		1.1053	1.1123	1.1192	1.1267	1.2
$ r=\frac{1}{\alpha_1} $	4.6880e-03	2.4428e-03	1.2576e-03	6.4230e-04	3.2627e-04
$ r=\frac{1}{\alpha_1} $		0.9404	0.9578	0.9694	0.9772	1.0
$ r=\frac{2-\alpha_{1}}{\alpha_{1}} $	4.9526e-03	2.3115e-03	1.0727e-03	4.9519e-04	2.2734e-04
$ r=\frac{2-\alpha_{1}}{\alpha_{1}} $		1.0994	1.1076	1.1152	1.1231	1.2

| Show Table

DownLoad: CSV

Table 7. Local $ H^1 $-norm errors and orders of convergence at $ t = 1 $ for Example 1 with $ \alpha_1 = 0.6 $, $ \alpha_{2} = 0.4 $.

	$ M=64 $	$ M=128 $	$ M=256 $	$ M=512 $	$ M=1024 $
$ r=1 $	3.2280e-03	1.5199e-03	7.1441e-04	3.3510e-04	1.5677e-04
$ r=1 $		1.0867	1.0891	1.0922	1.0959
$ r=2 $	4.3121e-03	2.0043e-03	9.2712e-04	4.2681e-04	1.9546e-04
$ r=2 $		1.1053	1.1123	1.1192	1.1267
$ r=\frac{1}{\alpha_1} $	3.7184e-03	1.7189e-03	7.9137e-04	3.6280e-04	1.6550e-04
$ r=\frac{1}{\alpha_1} $		1.1132	1.1190	1.1252	1.1323
$ r=\frac{2-\alpha_{1}}{\alpha_{1}} $	4.9526e-03	2.3115e-03	1.0727e-03	4.9519e-04	2.2734e-04
$ r=\frac{2-\alpha_{1}}{\alpha_{1}} $		1.0994	1.1076	1.1152	1.1231

| Show Table

DownLoad: CSV

Table 8. Global$ H^1 $-norm errors and orders of convergence for Example 1 with $ \alpha_1 = 0.8 $, $ \alpha_{2} = 0.6 $.

	$ M=64 $	$ M=128 $	$ M=256 $	$ M=512 $	$ M=1024 $	Theoretical order
$ r=1 $	7.1707e-03	4.2901e-03	2.5403e-03	1.4930e-03	8.7297e-04
$ r=1 $		0.7411	0.7560	0.7668	0.7743	0.8
$ r=2 $	1.9211e-03	7.5634e-04	2.9814e-04	1.1785e-04	4.6868e-05
$ r=2 $		1.3448	1.3430	1.3391	1.3302	1.2
$ r=\frac{1}{\alpha_1} $	3.7340e-03	1.9655e-03	1.0233e-03	5.2893e-04	2.7174e-04
$ r=\frac{1}{\alpha_1} $		0.9258	0.9417	0.9521	0.9608	1.0
$ r=\frac{2-\alpha_{1}}{\alpha_{1}} $	2.2479e-03	1.0660e-03	4.9903e-04	2.3116e-04	1.0617e-04
$ r=\frac{2-\alpha_{1}}{\alpha_{1}} $		1.0764	1.0950	1.1102	1.1225	1.2

| Show Table

DownLoad: CSV

Table 9. Local $ H^1 $-norm errors and orders of convergence at $ t = 1 $ for Example 1 with $ \alpha_1 = 0.8 $, $ \alpha_{2} = 0.6 $.

	$ M=64 $	$ M=128 $	$ M=256 $	$ M=512 $	$ M=1024 $
$ r=1 $	1.8988e-03	8.6835e-04	4.0348e-04	1.9000e-04	9.0390e-05
$ r=1 $		1.1287	1.1058	1.0865	1.0718
$ r=2 $	1.9211e-03	7.5634e-04	2.9814e-04	1.1785e-04	4.6664e-05
$ r=2 $		1.3448	1.3430	1.3391	1.3365
$ r=\frac{1}{\alpha_1} $	1.6142e-03	6.7476e-04	2.8389e-04	1.2013e-04	5.1015e-05
$ r=\frac{1}{\alpha_1} $		1.2584	1.2490	1.2407	1.2357
$ r=\frac{2-\alpha_{1}}{\alpha_{1}} $	1.6187e-03	6.5123e-04	2.6278e-04	1.0638e-04	4.3129e-05
$ r=\frac{2-\alpha_{1}}{\alpha_{1}} $		1.3136	1.3093	1.3046	1.3025

| Show Table

DownLoad: CSV

Figure 1. $ L^2 $-norm and $ H^1 $-norm errors and orders of convergence in spatial direction for Example 1 with $ \alpha_1 = 0.6 $, $ \alpha_{2} = 0.4 $ and $ r = 2 $.

DownLoad: Full-Size Img PowerPoint

Example 2. In Eq (1.1) take $ \Omega = [0, \pi]\times[0, \pi] $, $ T = 1 $, $ J = 2 $, $ b_1 = b_2 = 1 $. Let $ u_0(x, y) = \sin x\sin y $, the force term $ f(x, y, t) = 0 $.

The exact solution $ u(x, y, t) $ of Example 2 is unknown. We use the two-mesh principle in [4, p107] to calculate the convergence order of the numerical solution. Let $ u_{i, j}^{n} $ with $ 0 \leq i \leq N_1 $, $ 0 \leq j \leq N_2 $ and $ 0 \leq n \leq M $ be the solution computed by our L1-ADI Scheme 2.6. Then, consider a spatial mesh and the temporal mesh which is defined by $ 0 \leq i \leq N_1 $, $ 0 \leq j \leq N_2 $ and $ t_{n} = T(n /(2M)) {\text { for }} 0 \leq n \leq 2M $. We denote $ W_{h}^{n} $ with $ 0 \leq n \leq 2M $ as the computed solution on this mesh. Then, we define $ D_{h}^{n} = \left\|u_{h}^{n}-W_{h}^{2 n}\right\|_{H^1} $ as the $ H^1 $-norm of the two-mesh differences and the estimated rate of convergence is computed by $ \log _{2} (D_{h}^{n}/D_{h}^{2n}) $.

To test the temporal convergence of our scheme we set spatial partition parameters $ N_1 = N_2 = 1000 $. Tables 10, 12 and 14 present $ H^1 $-norm errors and convergence orders for Example 2 in the case of $ \alpha_{1} = 0.4, 0.6, 0.8 $ with $ \alpha_{2} = 0.2, 0.4, 0.6 $. The results show that the temporal convergence rates are $ O(M^{-\min \left\{ 2\alpha_{1}, r\alpha_1, 2-\alpha_1\right\}}) $, which, once again, confirm the sharpness of our theoretical analysis. We have also present the local $ H^1 $-norm errors and convergence orders at $ t = 1 $ for Example 2 in Tables 11, 13 and 15, which have better convergence rates than global errors.

Table 10. Global $ H^1 $-norm errors and orders of convergence for Example 2 with $ \alpha_1 = 0.4 $, $ \alpha_{2} = 0.2 $.

	$ M=64 $	$ M=128 $	$ M=256 $	$ M=512 $	$ M=1024 $	Theoretical order
$ r=1 $	4.2648e-02	3.5832e-02	2.9725e-02	2.4362e-02	1.9747e-02
$ r=1 $		0.2512	0.2696	0.2870	0.3030	0.4
$ r=2 $	1.7773e-02	1.1198e-02	6.8693e-03	4.1345e-03	2.4556e-03
$ r=2 $		0.6664	0.7050	0.7324	0.7517	0.8
$ r=\frac{1}{\alpha_1} $	9.6556e-03	5.1889e-03	2.8079e-03	1.4872e-03	9.0857e-04
$ r=\frac{1}{\alpha_1} $		0.8959	0.8859	0.9169	0.7109	0.8
$ r=\frac{2-\alpha_{1}}{\alpha_{1}} $	7.1020e-03	4.6925e-03	3.0434e-03	1.9441e-03	1.2265e-03
$ r=\frac{2-\alpha_{1}}{\alpha_{1}} $		0.5978	0.6247	0.6466	0.6646	0.8

| Show Table

DownLoad: CSV

Table 11. Local $ H^1 $-norm errors and orders of convergence at $ t = 1 $ for Example 2 with $ \alpha_1 = 0.4 $, $ \alpha_{2} = 0.2 $.

	$ M=64 $	$ M=128 $	$ M=256 $	$ M=512 $	$ M=1024 $
$ r=1 $	4.3792e-03	2.6480e-03	1.6045e-03	9.7164e-04	5.8717e-04
$ r=1 $		0.7258	0.7228	0.7236	0.7266
$ r=2 $	4.9902e-03	3.1856e-03	2.0192e-03	1.2690e-03	7.9068e-04
$ r=2 $		0.6475	0.6578	0.6701	0.6825
$ r=\frac{1}{\alpha_1} $	5.5605e-03	3.5950e-03	2.2975e-03	1.4521e-03	9.0857e-04
$ r=\frac{1}{\alpha_1} $		0.6292	0.6459	0.6619	0.6765
$ r=\frac{2-\alpha_{1}}{\alpha_{1}} $	7.1020e-03	4.6925e-03	3.0434e-03	1.9441e-03	1.2265e-03
$ r=\frac{2-\alpha_{1}}{\alpha_{1}} $		0.5978	0.6247	0.6466	0.6646

| Show Table

DownLoad: CSV

Table 12. Global $ H^1 $-norm errors and orders of convergence for Example 2 with $ \alpha_1 = 0.6 $, $ \alpha_{2} = 0.4 $.

	$ M=64 $	$ M=128 $	$ M=256 $	$ M=512 $	$ M=1024 $	Theoretical order
$ r=1 $	2.5383e-02	1.7829e-02	1.2380e-02	8.5185e-03	5.8193e-03
$ r=1 $		0.5096	0.5262	0.5394	0.5498	0.6
$ r=2 $	5.8541e-03	2.8312e-03	1.3301e-03	6.1587e-04	2.8032e-04
$ r=2 $		1.0480	1.0899	1.1108	1.1356	1.2
$ r=\frac{1}{\alpha_1} $	9.0483e-03	4.9340e-03	2.6263e-03	1.3755e-03	7.1241e-04
$ r=\frac{1}{\alpha_1} $		0.8749	0.9097	0.9331	0.9492	1.0
$ r=\frac{2-\alpha_{1}}{\alpha_{1}} $	4.1839e-03	1.8190e-03	7.6784e-04	3.1629e-04	1.4124e-04
$ r=\frac{2-\alpha_{1}}{\alpha_{1}} $		1.2017	1.2443	1.2796	1.1631	1.2

| Show Table

DownLoad: CSV

Table 13. Local $ H^1 $-norm errors and orders of convergence at $ t = 1 $ for Example 2 with $ \alpha_1 = 0.6 $, $ \alpha_{2} = 0.4 $.

	$ M=64 $	$ M=128 $	$ M=256 $	$ M=512 $	$ M=1024 $
$ r=1 $	3.0946e-03	1.4921e-03	7.2117e-04	3.4892e-04	1.6888e-04
$ r=1 $		1.0524	1.0489	1.0475	1.0469
$ r=2 $	2.8991e-03	1.3174e-03	5.9939e-04	2.7285e-04	1.2417e-04
$ r=2 $		1.1379	1.1361	1.1354	1.1358
$ r=\frac{1}{\alpha_1} $	2.6705e-03	1.2034e-03	5.4295e-04	2.4509e-04	1.1062e-04
$ r=\frac{1}{\alpha_1} $		1.1500	1.1482	1.1475	1.1477
$ r=\frac{2-\alpha_{1}}{\alpha_{1}} $	3.2093e-03	1.4709e-03	6.7404e-04	3.0872e-04	1.4124e-04
$ r=\frac{2-\alpha_{1}}{\alpha_{1}} $		1.1256	1.1258	1.1265	1.1282

| Show Table

DownLoad: CSV

Table 14. Global $ H^1 $-norm errors and orders of convergence for Example 2 with $ \alpha_1 = 0.8 $, $ \alpha_{2} = 0.6 $.

	$ M=64 $	$ M=128 $	$ M=256 $	$ M=512 $	$ M=1024 $	Theoretical order
$ r=1 $	1.1164e-02	6.6335e-03	3.9421e-03	2.3268e-03	1.3681e-03
$ r=1 $		0.7511	0.7508	0.7606	0.7662	0.8
$ r=2 $	3.9574e-03	1.7177e-03	7.3884e-04	3.1639e-04	1.3526e-04
$ r=2 $		1.2041	1.2171	1.2235	1.2260	1.2
$ r=\frac{1}{\alpha_1} $	7.2913e-03	3.8726e-03	2.0378e-03	1.0644e-03	5.5259e-04
$ r=\frac{1}{\alpha_1} $		0.9129	0.9263	0.9370	0.9457	1.0
$ r=\frac{2-\alpha_{1}}{\alpha_{1}} $	5.2772e-03	2.5404e-03	1.2056e-03	5.6564e-04	2.6297e-04
$ r=\frac{2-\alpha_{1}}{\alpha_{1}} $		1.0547	1.0754	1.0918	1.1050	1.2

| Show Table

DownLoad: CSV

Table 15. Local $ H^1 $-norm errors and orders of convergence at $ t = 1 $ for Example 2 with $ \alpha_1 = 0.8 $, $ \alpha_{2} = 0.6 $.

	$ M=64 $	$ M=128 $	$ M=256 $	$ M=512 $	$ M=1024 $
$ r=1 $	3.5614e-03	1.6739e-03	7.9585e-04	3.8211e-04	1.8495e-04
$ r=1 $		1.0892	1.0727	1.0585	1.0468
$ r=2 $	3.3083e-03	1.3585e-03	5.5756e-04	2.2938e-04	9.4727e-05
$ r=2 $		1.2840	1.2848	1.2814	1.2759
$ r=\frac{1}{\alpha_1} $	3.0927e-03	1.3497e-03	5.9116e-04	2.5979e-04	1.1450e-04
$ r=\frac{1}{\alpha_1} $		1.1962	1.1910	1.1862	1.1820
$ r=\frac{2-\alpha_{1}}{\alpha_{1}} $	3.0053e-03	1.2639e-03	5.3187e-04	2.2422e-04	9.4744e-05
$ r=\frac{2-\alpha_{1}}{\alpha_{1}} $		1.2496	1.2487	1.2462	1.2428

| Show Table

DownLoad: CSV

6. Conclusion

A fully discrete L1-ADI scheme is investigated for the initial-boundary problem of a multi-term time-fractional diffusion equation. Stability and convergence of the fully discrete L1-ADI scheme are rigorously established. Both $ L^2 $-norm and $ H^1 $-norm error estimates of the fully discrete L1-ADI scheme are obtained, and they are $ \alpha $-robust. Numerical experiments are given to illustrate the sharpness of the theoretical analysis. It should be noted that since the computational cost of numerical methods for time-fractional PDEs will be very time-consuming. One can also use the fast and parallel numerical methods such as ^[5,24,32] for accelerating the proposed method, which will be the focus of our future work.

Use of AI tools declaration

The authors declare they have not used Artificial Intelligence (AI) tools in the creation of this article.

Acknowledgments

The research is supported in part by the National Natural Science Foundation of China under Grant 11801026, and Fundamental Research Funds for the Central Universities (No. 202264006).

Conflict of interest

The authors declare there is no conflict of interest.

References

[1]	The Florida Center of Excellence in Drug Discovery and Innovation (CDDI) is described at: http://www.research.usf.edu/cddi/drugdiscovery/resources.asp.
[2]	Huang YM, Amsler AO, McClintock JB, et al. (2007) Patterns of gammarid amphipod abundance and species composition associated with dominant subtidal macroalgae along the western Antarctic Peninsula. Polar Biol 30: 1417-1430. doi: 10.1007/s00300-007-0303-1
[3]	Santiago DN, Pevzner Y, Durand AA, et al. (2012) Virtual Target Screening: Validation Using Kinase Inhibitors. J Chem Inf Model 52: 2192-2203. doi: 10.1021/ci300073m
[4]	Irwin JJ, Sterling T, Mysinger MM, et al. (2012) ZINC: A Free Tool to Discover Chemistry for Biology. J Chem Info Model 52:1757-1768. doi: 10.1021/ci3001277
[5]	Gerwick WH, Moore BS (2012) Lessons from the past and charting the future of marine natural products drug discovery and chemical biology. Chem Biol 19:85-98. doi: 10.1016/j.chembiol.2011.12.014
[6]	Swinney DC, Anthony J (2011) How were new medicines discovered? Nat Rev Drug Discov 10: 507-519. doi: 10.1038/nrd3480
[7]	Newman DJ, Cragg GM (2012) Natural Products as Sources of New Drugs over the 30 Years from 1981 to 2010. J Nat Prod 75: 311-335. doi: 10.1021/np200906s
[8]	Mishra BB, Tiwari VK (2011) Natural products: An evolving role in future drug discovery. Eur J Med Chem 46: 4769-4807. doi: 10.1016/j.ejmech.2011.07.057
[9]	Bachmann BO, Van Lanen SG, Baltz RH (2014) Microbial genome mining for accelerated natural products discovery: is a renaissance in the making? J Indust Microbiol Biotech 41: 175-184. doi: 10.1007/s10295-013-1389-9
[10]	Williams P, Sorribasa A, Howes MR (2011) Natural products as a source of Alzheimer's drug leads. Nat Prod Rep 28: 48-77. doi: 10.1039/C0NP00027B
[11]	Kuete V, Alibert-Franco S, Eyong KO, et al. (2011) Antibacterial activity of some natural products against bacteria expressing a multidrug-resistant phenotype. Int J Antimicrob Agents 37: 156-161. doi: 10.1016/j.ijantimicag.2010.10.020
[12]	Cragg GM, Newman DJ (2013) Natural products: A continuing source of novel drug leads. Biochim Biophys Acta 1830: 3670-3695. doi: 10.1016/j.bbagen.2013.02.008
[13]	Dayan FE, Owens DK, Duke SO (2012) Rationale for a natural products approach to herbicide discovery. Pest Mgmt Sci 68: 519-528. doi: 10.1002/ps.2332
[14]	Bugni TS, Harper MK, McCulloch MWB, et al. (2008) Fractionated marine invertebrate extract libraries for drug discovery. Molecules 13: 1372-1383. doi: 10.3390/molecules13061372
[15]	Molinski TF (2010) NMR of natural products at the “nanomole-scale”. Nat Prod Rep 27: 321-329. doi: 10.1039/b920545b
[16]	Nicolaou KC, Snyder SA (2005) Chasing molecules that were never there: misassigned natural products and the role of chemical synthesis in modern structure elucidation. Angew. Chem Int Ed Engl 44: 1012-1044. doi: 10.1002/anie.200460864
[17]	Penn K, Jenkins C, Nett M, et al. (2009) Genomic islands link secondary metabolism to functional adaptation in marine Actinobacteria. ISME J 3: 1193-1203. doi: 10.1038/ismej.2009.58
[18]	Lipinski CA, Lombardo F, Dominy BW, et al. (2001) Experimental and computational approaches to estimate solubility and permeability in drug discovery and development settings. Adv Drug Deliv Rev 46: 3-26. doi: 10.1016/S0169-409X(00)00129-0
[19]	Ursu O, Rayan A, Goldblum A, et al. (2011) Understanding drug-likeness. WIREs: Comp Mol Sci 1:760-781. doi: 10.1002/wcms.52
[20]	Bickerton GR, Paoliuni GV, Besnard J, et al. (2012) Quantifying the chemical beauty of drugs. Nat Chem 4: 90-98. doi: 10.1038/nchem.1243
[21]	Matter H, Sotriffer C (2011) Applications and success stories in virtual screening. In: Virtual screening: principles, challenges, and practical guidelines. Wiley-VCH Verlag GmbH & Co. Ch. 12: 319-358.
[22]	Vangrevelinghe E, Zimmermann K, Schoepfer J, et al. (2003) Discovery of a Potent and Selective Protein Kinase CK2 Inhibitor by High-Throughput Docking. J Med Chem 46: 2656-2662. doi: 10.1021/jm030827e
[23]	Ewing TA, Makino S, Skillman AG, et al. (2001) DOCK 4.0: search strategies for automated molecular docking of flexible molecule databases. J Comp Mol Des 15: 411-428.
[24]	Friesner RA, Banks JL, Murphy RB, et al. (2004) Glide: New approach for rapid, accurate docking and scoring. 1. Method and Assessment of Docking Accuracy. J Med Chem 47: 1739-1749.
[25]	26. Brooks WH, McCloskey DE, Daniel KE, et al. (2007) In Silico Chemical Library Screening and Experimental Validation of a Novel 9-Aminoacridine Based Lead-Inhibitor of Human S-Adenosylmethionine Decarboxylase. J Chem Info Model 47: 1897-1905. doi: 10.1021/ci700005t
[26]	27. Tolbert WD, Ekstrom JL, Mathews II, et al. (2001) The structural basis for substrate specificity and inhibition of human S-adenosylmethionine decarboxylase. Biochemistry 40: 9484-9494. doi: 10.1021/bi010735w
[27]	28. Berman HM, Westbrook J, Feng Z, et al. (2000) The Protein Data Bank. Nucl Acids Res 28: 235-242. doi: 10.1093/nar/28.1.235
[28]	29. Hui-Fang L, Qing S, Jian Z, et al. (2010) Evaluation of various inverse docking schemes in multiple targets identification. J Mol Graph Model 29: 326-330. doi: 10.1016/j.jmgm.2010.09.004
[29]	30. Grinter SZ, Liang Y, Huang SY, et al. (2011) An inverse docking approach for identifying new potential anti-cancer targets. J Mol Graph Model 795-799. doi: 10.1016/j.jmgm.2011.01.002
[30]	31. Do QT, Lamy C, Renimel I, et al. (2007) Reverse Pharmacognosy: Identifying Biological Properties for Plants by Means of their Molecule Constituents: Application to Meranzin. Planta Med 73: 1235-1240. doi: 10.1055/s-2007-990216
[31]	32. Li YY, An J, Jones SJM (2006) A Large-Scale Computational Approach to Drug Repositioning. Genome Info 17: 239-247.
[32]	37. Miller BT, Singh RP, Klauda JB, et al. (2008) CHARMMing: a new, flexible web portal for CHARMM. J Chem Info Model 48: 1920-1929. doi: 10.1021/ci800133b
[33]	38. Wang Y, Xiao J, Suzek TO, et al. (2012) PubChem's BioAssay Database. Nucl Acids Res 40: D400-12. doi: 10.1093/nar/gkr1132
[34]	39. Bolton E, Wang Y, Thiessen PA, et al. (2008) PubChem: Integrated Platform of Small Molecules and Biological Activities. Chapter 12 in Annual Reports in Computational Chemistry, Volume 4, ACS, Washington, DC.
[35]	42. Imhoff JF, Labes A, Wiese J (2011) Bio-mining the microbial treasures of the ocean: New natural products. Biotech Adv 29: 468-482. doi: 10.1016/j.biotechadv.2011.03.001
[36]	43. Liu X, Ashforth E, Ren B, et al. (2010) Bioprospecting microbial natural product libraries from the marine environment for drug discovery. J Antibiotics 63: 415-422. doi: 10.1038/ja.2010.56
[37]	44. Bauer RA, Wurst JM, Tan DS (2010) Expanding the range of “druggable” targets with natural product-based libraries: an academic perspective. Curr Op Chem Biol 14: 308-314. doi: 10.1016/j.cbpa.2010.02.001
[38]	45. Mayer AMS, Glaser KB, Cuevas C, et al. (2010) The odyssey of marine pharmaceuticals: a current pipeline perspective. Trends in Pharm Sci 31: 255-265. doi: 10.1016/j.tips.2010.02.005
[39]	46. Bennani YL (2011) Drug discovery in the next decade: innovation needed ASAP. Drug Disc Today 16:
[40]	47. Keiser MJ, Setola V, Irwin JJ, et al. (2009) Predicting new molecular targets for known drugs. Nature 462: 175-182. doi: 10.1038/nature08506
[41]	48. Reker D, Rodrigues T, Schneider P, et al. (2014) Identifying the macromolecular targets of de novo-designed chemical entities through self-organizing map consensus. Proc Natl Acad Sci USA 111: 4067-4072. doi: 10.1073/pnas.1320001111
[42]	49. Liu X, Vogt I, Hague T, et al. (2013) HitPick: A web server for hit identification and target prediction of chemical screenings. Bioinformatics 29: 1910-1912. doi: 10.1093/bioinformatics/btt303
[43]	50. Gong J, Cai C, Liu X, et al. (2013) ChemMapper: A versatile web server for exploring pharmacology and chemical structure association based on molecular 3D similarity method. Bioinformatics 29: 1827-1829. doi: 10.1093/bioinformatics/btt270
[44]	51. Iorio F, Bosotti R, Scachen E, et al. (2010) Discovery of drug mode of action and drug repositioning from transcriptional responses. Proc Nat Acad Sic USA 107: 14621-14626. doi: 10.1073/pnas.1000138107
[45]	52. Zhu X, Kim JL, Newcomb JR, et al. (1999) Structural analysis of the lymphocyte-specific kinase Lck in complex with non-selective and Src family selective kinase inhibitors. Struct Fold Des 7: 651-661. doi: 10.1016/S0969-2126(99)80086-0
[46]	53. Goto M, Miyahara I, Hirotsu K, et al. (2005) Structural determinants for branched-chain aminotransferase isozyme-specific inhibition by the anticonvulsant drug gabapentin. J Biol Chem 280: 37237256. doi: 10.1074/jbc.M506486200
[47]	55. Halgren T (2009) Identifying and characterizing binding sites and assessing druggability. J Chem Inf Model 49:377-389. doi: 10.1021/ci800324m

This article has been cited by:

1.	Xin Shen, Xuehua Yang, Haixiang Zhang, The High-Order ADI Difference Method and Extrapolation Method for Solving the Two-Dimensional Nonlinear Parabolic Evolution Equations, 2024, 12, 2227-7390, 3469, 10.3390/math12223469
2.	Wang Xiao, Xuehua Yang, Ziyi Zhou, Pointwise-in-time $ \alpha $-robust error estimate of the ADI difference scheme for three-dimensional fractional subdiffusion equations with variable coefficients, 2024, 16, 2836-3310, 53, 10.3934/cam.2024003
3.	Adel Al-rabtah, Salah Abuasad, Effective Modified Fractional Reduced Differential Transform Method for Solving Multi-Term Time-Fractional Wave-Diffusion Equations, 2023, 15, 2073-8994, 1721, 10.3390/sym15091721
4.	Pradip Roul, Trishna Kumari, Sameer N. Khandagale, A graded mesh technique for numerical approximation of a multi-term Caputo time-fractional Fokker-Planck equation in 2D space, 2025, 178, 08981221, 92, 10.1016/j.camwa.2024.11.031

Reader Comments

Your name:*

Email:*
© 2014 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)

通讯作者: 陈斌, bchen63@163.com

1.
沈阳化工大学材料科学与工程学院沈阳 110142

AIMS Molecular Science

0.7

Metrics

Article views(5852) PDF downloads(986) Cited by(1)

Preview PDF

Download XML

Export Citation

Article outline

Show full outline

Figures and Tables

Figures(2) / Tables(2)

AIMS Molecular Science

Virtual target screening to rapidly identify potential protein targets of natural products in drug discovery

Related Papers:

Abstract

1. Introduction

2. Fully discrete L1-ADI scheme

3. Analysis of stability and convergence

4. $ H^1 $-norm convergence of L1-ADI scheme

5. Numerical experiments

6. Conclusion

Use of AI tools declaration

Acknowledgments

Conflict of interest

References

This article has been cited by:

Reader Comments

通讯作者: 陈斌, bchen63@163.com

Metrics

Figures and Tables

Other Articles By Authors

Catalog

AIMS Molecular Science

Virtual target screening to rapidly identify potential protein targets of natural products in drug discovery

Related Papers:

Abstract

1. Introduction

2. Fully discrete L1-ADI scheme

3. Analysis of stability and convergence

4. $ H^1 $-norm convergence of L1-ADI scheme

5. Numerical experiments

6. Conclusion

Use of AI tools declaration

Acknowledgments

Conflict of interest

References

This article has been cited by:

Reader Comments

通讯作者: 陈斌, bchen63@163.com

Metrics

Figures and Tables

Other Articles By Authors

Related pages

Tools

Export File

Citation

Format

Content

Catalog