Hindawi Journal of Function Spaces Volume 2017, Article ID 3679526, 8 pages https://doi.org/10.1155/2017/3679526
Research Article A Compact Difference Scheme for Solving Fractional Neutral Parabolic Differential Equation with Proportional Delay Wei Gu,1 Yanli Zhou,2 and Xiangyu Ge1 1
School of Statistics & Mathematics, Zhongnan University of Economics and Law, Wuhan 430073, China School of Finance, Zhongnan University of Economics and Law, Wuhan 430073, China
2
Correspondence should be addressed to Yanli Zhou;
[email protected] and Xiangyu Ge; xiangyu
[email protected] Received 5 July 2017; Accepted 17 September 2017; Published 18 October 2017 Academic Editor: Xinguang Zhang Copyright ยฉ 2017 Wei Gu et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. A linearized compact finite difference scheme is constructed for solving the fractional neutral parabolic differential equation with proportional delay. By the energy method, the unconditional stability of the scheme is proved, and the convergence order of the scheme is proved to be ๐(๐2โ๐ผ + โ4 ). A numerical test is also conducted to validate the accuracy and efficiency of the numerical algorithm.
1. Introduction
๐ค (0, ๐ก) = ๐พ๐ค (๐ก) ,
In the past few years, more and more scholars have been attracted to the research of delay partial differential equations (DPDEs) [1โ3]. However, most DPDEs have no exact solutions. Constructing efficient numerical methods for DPDEs is of great importance [4โ7]. For details on numerically solving neutral delay parabolic differential equations (NDPDEs), the reader is referred to [5, 8]. Recently, fractional delay partial differential equations have been of great interest due to their application in automatic control, population dynamics, economics, and so forth [9, 10]. For details on numerical solutions to fractional delay partial differential equations, we refer the reader to [11, 12]. The work in [11] considers the numerical method without theoretical analysis, and the work in [12] considers the numerical method for a type of semilinear fractional partial differential equation with time delay. In this paper, we consider the following fractional neutral parabolic differential equation with proportional delay:
๐ค (1, ๐ก) = ๐ฝ๐ค (๐ก) ,
2 ๐2 ๐ค (๐ฅ, ๐ก) ๐ ๐ค (๐ฅ, ๐๐ก) ๐๐ผ ๐ค = ๐ท( + ) ๐ผ 2 ๐๐ก ๐๐ฅ ๐๐ฅ2
๐ก โ (0, ๐] , (1) where ๐ท > 0 is a constant, 0 < ๐ < 1. Let ๐ข(๐ฅ, ๐ก) = ๐ค(๐ฅ, ๐๐ก ) for ๐ก โฅ ๐ก0 + ln(๐), where ๐ก0 โฅ 0. Then, ๐ข(๐ฅ, ๐ก) satisfies the following equation: ๐๐ผ ๐ข ๐2 ๐ข (๐ฅ, ๐ก) ๐2 ๐ข (๐ฅ, ๐ก โ ๐ ) = ๐ท ( + ) ๐๐ก๐ผ ๐๐ฅ2 ๐๐ฅ2 + ๐๐ค (๐ข (๐ฅ, ๐ก) , ๐ข (๐ฅ, ๐ก โ ๐ ) , ๐ฅ, ๐๐ก ) , (๐ฅ, ๐ก) โ (0, 1) ร (๐ก0 , ๐] , ๐ข (๐ฅ, ๐ก) = ๐๐ค (๐ฅ, ๐๐ก ) , ๐ข (0, ๐ก) = ๐พ๐ค (๐๐ก ) , ๐ข (1, ๐ก) = ๐ฝ๐ค (๐๐ก ) , ๐ก โ (๐ก0 , ๐] ,
+ ๐๐ค (๐ค (๐ฅ, ๐ก) , ๐ค (๐ฅ, ๐๐ก) , ๐ฅ, ๐ก) , (๐ฅ, ๐ก) โ (0, 1) ร [0, ๐] ,
๐ฅ โ [0, 1] , ๐ก โ [โ๐ , ๐ก0 ] ,
where ๐ = โ ln(๐).
(2)
2
Journal of Function Spaces
For simplicity, we consider the following fractional neutral parabolic differential equation with delay instead of (1): ๐๐ผ ๐ข ๐2 ๐ข (๐ฅ, ๐ก) ๐2 ๐ข (๐ฅ, ๐ก โ ๐ ) = ๐ท ( + ) ๐๐ก๐ผ ๐๐ฅ2 ๐๐ฅ2
๐ = ๐ฟ๐ฅ V๐+1/2
(3)
+ ๐ (๐ข (๐ฅ, ๐ก) , ๐ข (๐ฅ, ๐ก โ ๐ ) , ๐ฅ, ๐ก) , (๐ฅ, ๐ก) โ (0, 1) ร (0, ๐] , ๐ข (๐ฅ, ๐ก) = ๐ (๐ฅ, ๐ก) ,
be the grid function space defined on ฮฉโ๐ . The following notations are used:
๐ฅ โ [0, 1] , ๐ก โ [โ๐ , 0] ,
(4)
๐ข (0, ๐ก) = ๐พ (๐ก) , ๐ข (1, ๐ก) = ๐ฝ (๐ก) ,
(5) ๐ก โ (0, ๐] ,
where ๐ > 0 is a constant delay term. Time fractional partial derivative (๐๐ผ ๐ข/๐๐ก๐ผ ) (0 < ๐ผ < 1) is defined in the Caputo sense by the following: ๐ก ๐๐ผ ๐ข 1 โ๐ผ ๐๐ข (๐ฅ, ๐) = ๐๐, โซ (๐ก โ ๐) ๐๐ก๐ผ ฮ (1 โ ๐ผ) 0 ๐๐
(6)
where ฮ( ) is the Gamma function. In this paper, a linearized compact finite difference scheme is constructed for solving (3)โ(5). By the energy method, the unconditional stability of the scheme is then proved, and the convergence order of the scheme is proved to be ๐(๐2โ๐ผ +โ4 ). A numerical test is also conducted to validate the accuracy and efficiency of the numerical algorithm. The rest of the paper is organized as follows. In Section 2, a compact difference scheme is constructed to solve (3)โ(5). Section 3 considers the solvability, convergence, and stability of the provided difference scheme. In Section 4, a numerical test is presented to illustrate the validity of the theoretical results. Section 5 gives a brief conclusion of this paper.
2. The Construction of the Compact Difference Scheme
๐ฟ๐ฅ2 V๐๐ =
๐ ๐ โ 2V๐๐ + V๐โ1 V๐+1 , โ2
AV๐๐ =
1 ๐ ๐ ), (V + 10V๐๐ + V๐+1 12 ๐โ1
๐ฟ๐ก๐ผ V๐ =
โ๐ผ
(9)
๐โ1
๐ [๐0 V๐ โ โ (๐๐โ๐โ1 โ ๐๐โ๐ ) V๐ ฮ (2 โ ๐ผ) ๐=1 [
โ ๐๐โ1 V0 ] , ] where ๐๐ = (๐ + 1)1โ๐ผ โ ๐1โ๐ผ , ๐ โฅ 0. For the time fractional derivative, we have the following lemma. Lemma 1 (see [13]). Suppose 0 < ๐ผ < 1, ๐ฆ โ ๐ถ2 [0, ๐ก๐ ]; it holds that ๓ตจ๓ตจ ๓ตจ๓ตจ ๐ก๐ ๐ฆ๓ธ (๐ ) ๐๐ 1 ๐โ๐ผ [ ๓ตจ๓ตจ ๐0 ๐ฆ (๐ก๐ ) โซ ๓ตจ๓ตจ ๐ผ โ ๓ตจ๓ตจ ฮ (1 โ ๐ผ) 0 (๐ก๐ โ ๐ ) ฮ (2 โ ๐ผ) ๓ตจ๓ตจ [ ๓ตจ๓ตจ ๓ตจ๓ตจ ๓ตจ โ โ (๐๐โ๐โ1 โ ๐๐โ๐ ) ๐ฆ (๐ก๐ ) โ ๐๐โ1 ๐ฆ (๐ก0 )]๓ตจ๓ตจ๓ตจ ๓ตจ๓ตจ ๐=1 ]๓ตจ๓ตจ ๐โ1
โค
(10)
1 โ ๐ผ 22โ๐ผ 1 [ + โ (1 + 2โ๐ผ )] ฮ (2 โ ๐ผ) 12 2โ๐ผ
๓ตจ ๓ตจ โ
max ๓ตจ๓ตจ๓ตจ๓ตจ๐ฆ๓ธ ๓ธ (๐ก)๓ตจ๓ตจ๓ตจ๓ตจ ๐2โ๐ผ , 0โค๐กโค๐ก๐ and ๐๐ satisfies the following lemma.
Throughout this paper, assume ๐ข(๐ฅ, ๐ก) โ ๐ถ6,2 ([0, 1] ร (0, ๐]). Function ๐(๐ข(๐ฅ, ๐ก), ๐ข(๐ฅ, ๐ก โ ๐ ), ๐ฅ, ๐ก) is sufficiently smooth and satisfies ๓ตจ๓ตจ๓ตจ๐ (๐ + ๐1 , ] + ๐2 , ๐ฅ, ๐ก) โ ๐ (๐, ], ๐ฅ, ๐ก)๓ตจ๓ตจ๓ตจ ๓ตจ ๓ตจ (7) ๓ตจ๓ตจ ๓ตจ๓ตจ ๓ตจ๓ตจ ๓ตจ๓ตจ โค ๐0 ๓ตจ๓ตจ๐1 ๓ตจ๓ตจ + ๐1 ๓ตจ๓ตจ๐2 ๓ตจ๓ตจ , where ๐1 , ๐2 are arbitrary real numbers and ๐0 and ๐1 are positive constants. First, let ๐ and ๐ be two positive integers; then, we take โ = 1/๐, ๐ = ๐ /๐ (๐ > 0 is a positive integer), ๐ฅ๐ = ๐โ, ๐ก๐ = ๐๐. Define ฮฉโ๐ = ฮฉโ ร ฮฉ๐ , where ฮฉโ = {๐ฅ๐ | 0 โค ๐ โค ๐}, ฮฉ๐ = {๐ก๐ | โ๐ โค ๐ โค ๐}, ๐ = [๐/๐]. Denote ๐๐๐ = ๐ข(๐ฅ๐ , ๐ก๐ ), 0 โค ๐ โค ๐, โ๐ โค ๐ โค ๐ throughout this paper. Let W = {V๐๐ | 0 โค ๐ โค ๐, โ๐ โค ๐ โค ๐}
๐ โ V๐๐ V๐+1 , โ
(8)
Lemma 2 (see [14]). Assume 0 < ๐ผ < 1; then, it holds that (1) ๐๐ decreases monotonically as ๐ increases, and 0 < ๐๐ โค 1; (2) ๐0 = 1, โ๐โ1 ๐=1 (๐๐โ๐โ1 โ ๐๐โ๐ ) = ๐0 โ ๐๐โ1 . Lemma 3 (see [5]). Suppose ๐ฆ(๐ฅ) โ ๐ถ6 [๐ฅ๐โ1 , ๐ฅ๐+1 ]; then, one has 1 [๐ฆ๓ธ ๓ธ (๐ฅ๐โ1 ) + 10๐ฆ๓ธ ๓ธ (๐ฅ๐ ) + ๐ฆ๓ธ ๓ธ (๐ฅ๐+1 )] 12 โ =
1 [๐ฆ (๐ฅ๐โ1 ) โ 2๐ฆ (๐ฅ๐ ) + ๐ฆ (๐ฅ๐+1 )] โ2
โ4 (6) ๐ฆ (๐๐ ) , 240
where ๐๐ โ (๐ฅ๐โ1 , ๐ฅ๐+1 ).
(11)
Journal of Function Spaces
3
Considering (3) at the point (๐ฅ๐ , ๐ก๐ ), we have
where
๐๐ผ ๐ข ๐2 ๐ข ๐2 ๐ข (๐ฅ , ๐ก ) = ๐ท ( (๐ฅ , ๐ก ) + (๐ฅ , ๐ก โ ๐ )) ๐ ๐ ๐ ๐ ๐๐ก๐ผ ๐๐ฅ2 ๐๐ฅ2 ๐ ๐ + ๐ (๐ข (๐ฅ๐ , ๐ก๐ ) , ๐ข (๐ฅ๐ , ๐ก๐ โ ๐ ) , ๐ฅ๐ , ๐ก๐ ) ,
๐ + ๐
๐๐ = A๐
0๐
From Lemma 1, we obtain ๐๐ผ ๐ข (๐ฅ , ๐ก ) = ๐ฟ๐ก๐ผ ๐๐๐ + ๐๐๐ , ๐๐ก๐ผ ๐ ๐
(13)
where ๐๐๐ =
1 1 โ ๐ผ 22โ๐ผ [ + โ (1 + 2โ๐ผ )] ฮ (2 โ ๐ผ) 12 2โ๐ผ ๓ตจ๓ตจ ๐2 ๐ข (๐ฅ , ๐ก ) ๓ตจ๓ตจ ๓ตจ ๐ ๐ ๓ตจ๓ตจ 2โ๐ผ ๓ตจ๓ตจ ๐ . โ
max ๓ตจ๓ตจ๓ตจ๓ตจ 2 ๓ตจ๓ตจ 0โค๐กโค๐ก๐ ๓ตจ ๐๐ก ๓ตจ ๓ตจ
(14)
Replacing ๐๐๐ by ๐ข๐๐ in (19), (22), and (23) and omitting ๐
๐๐ , we can obtain the following compact difference scheme: (15)
2
where ๐๐ โ (๐ก๐โ2 , ๐ก๐ ), ๓ฐ๐๐ in between ๐ข(๐ฅ๐ , ๐ก๐ ) and 2๐๐๐โ1 โ๐๐๐โ2 . Substituting (13) and (15) into (12) and applying the operator A on both sides of (12), we obtain ๐2 ๐ข ๐2 ๐ข A๐ฟ๐ก๐ผ ๐๐๐ = ๐ท (A 2 (๐ฅ๐ , ๐ก๐ ) + A 2 (๐ฅ๐ , ๐ก๐ โ ๐ )) ๐๐ฅ ๐๐ฅ
(17)
๐2 ๐ข โ4 ๐6 ๐ข ๐ 2 ๐ (๐ฅ , ๐ก ) = ๐ฟ ๐ + (๐ , ๐ก ) , ๐ฅ ๐ ๐๐ฅ2 ๐ ๐ 240 ๐๐ฅ6 ๐ ๐
๐ข0๐ = ๐พ (๐ก๐ ) , ๐ = ๐ฝ (๐ก๐ ) , ๐ข๐
(26) 1 โค ๐ โค ๐.
Define the following grid function space on ฮฉโ : Vโ,0 = {๐ข | ๐ข = (๐ข0 , ๐ข1 , . . . , ๐ข๐) , ๐ข0 = ๐ข๐ = 0} .
(27)
๐โ1
๐๐๐ โ (๐ฅ๐โ1 , ๐ฅ๐+1 ) , โ4 ๐6 ๐ข ๐โ๐ ๐2 ๐ข (๐ , ๐ก๐โ๐ ) , A 2 (๐ฅ๐ , ๐ก๐ โ ๐ ) = ๐ฟ๐ฅ2 ๐๐๐โ๐ + ๐๐ฅ 240 ๐๐ฅ6 ๐
(๐ข, V) = โ โ ๐ข๐ V๐ , ๐=1
(18)
๓ตจ ๓ตจ โ๐ขโ = โ(๐ข, ๐ข), โ๐ขโโ = max ๓ตจ๓ตจ๓ตจ๐ข๐ ๓ตจ๓ตจ๓ตจ , 1โค๐โค๐โ1 ๐โ1
โจ๐ฟ๐ฅ ๐ข, ๐ฟ๐ฅ Vโฉ = โ โ (๐ฟ๐ฅ ๐ข๐+1/2 ) (๐ฟ๐ฅ V๐+1/2 ) ,
๐๐๐โ๐ โ (๐ฅ๐โ1 , ๐ฅ๐+1 ) .
๐=0
Substituting (18) into (16), we have
๓ตจ๓ตจ ๓ตจ๓ตจ ๓ตจ๓ตจ๐ฟ๐ฅ ๐ข๓ตจ๓ตจ1 = โโจ๐ฟ๐ฅ ๐ข, ๐ฟ๐ฅ ๐ขโฉ.
A๐ฟ๐ก๐ผ ๐๐๐ = ๐ท (๐ฟ๐ฅ2 ๐๐๐ + ๐ฟ๐ฅ2 ๐๐๐โ๐ ) โ
(25)
If ๐ข, V โ Vโ,0 , we introduce the following inner products and corresponding norms:
From Lemma 3 and Taylor expansion, we have
+
(24)
3. The Solvability, Convergence, and Stability of the Difference Scheme
where
๐๐๐โ2 , ๐๐๐โ๐ , ๐ฅ๐ , ๐ก๐ )
๐ข๐๐ = ๐ (๐ฅ๐ , ๐ก๐ ) , 0 โค ๐ โค ๐, โ๐ โค ๐ โค 0,
(16)
๐ + A๐ (2๐๐๐โ1 โ ๐๐๐โ2 , ๐๐๐โ๐ , ๐ฅ๐ , ๐ก๐ ) + A๐
0๐ ,
๐2 ๐ข (๐ฅ , ๐๐ ) ๐๐ (๓ฐ๐๐ , ๐๐๐โ๐ , ๐ฅ๐ , ๐ก๐ ) . ๐๐ก2 ๐
A๐ฟ๐ก๐ผ ๐ข๐๐ = ๐ท (๐ฟ๐ฅ2 ๐ข๐๐ + ๐ฟ๐ฅ2 ๐ข๐๐โ๐ ) + A๐ (2๐ข๐๐โ1 โ ๐ข๐๐โ2 , ๐ข๐๐โ๐ , ๐ฅ๐ , ๐ก๐ ) ,
๐ข + ๐ 2 (๐ฅ๐ , ๐๐ ) ๐๐ (๓ฐ๐๐ , ๐๐๐โ๐ , ๐ฅ๐ , ๐ก๐ ) , ๐๐ก
A๐ (2๐๐๐โ1
(23) 1 โค ๐ โค ๐.
= ๐ (2๐๐๐โ1 โ ๐๐๐โ2 , ๐๐๐โ๐ , ๐ฅ๐ , ๐ก๐ )
A
(22)
๐0๐ = ๐พ (๐ก๐ ) , ๐ = ๐ฝ (๐ก๐ ) , ๐๐
๐ (๐ข (๐ฅ๐ , ๐ก๐ ) , ๐ข (๐ฅ๐ , ๐ก๐ โ ๐ ) , ๐ฅ๐ , ๐ก๐ )
๐ = ๐๐๐ + ๐2 ๐
0๐
(20)
Noticing ๐ข(๐ฅ, ๐ก) โ ๐ถ6,2 ([0, 1] ร (0, ๐]) and (7), we can easily obtain ๓ตจ๓ตจ๓ตจ๐
๐ ๓ตจ๓ตจ๓ตจ โค ๐ถ (๐2โ๐ผ + โ4 ) , 0 โค ๐ โค ๐, 1 โค ๐ โค ๐. (21) ๓ตจ๓ตจ ๐ ๓ตจ๓ตจ ๐
Discretizing the initial and boundary conditions of (4) and (5), we obtain ๐๐๐ = ๐ (๐ฅ๐ , ๐ก๐ ) , 0 โค ๐ โค ๐, โ๐ โค ๐ โค 0,
From Taylor expansion, we have
2๐
โ4 ๐6 ๐ข ๐โ๐ (๐ , ๐ก๐โ๐ ) . + 240 ๐๐ฅ6 ๐
(12)
0 โค ๐ โค ๐, 0 โค ๐ โค ๐.
โ4 ๐6 ๐ข ๐ (๐ , ๐ก ) 240 ๐๐ฅ6 ๐ ๐
+
๐
๐๐ ,
(19)
It is easy to obtain the following lemma. Lemma 4 (see [12]). โ๐ข โ Vโ,0 , one has โA๐ขโ2 โค โ๐ขโ2 .
(28)
4
Journal of Function Spaces
Lemma 5 (see [15]). โ๐ข โ Vโ,0 , one has
where ๐(๐ฅ, ๐ก) is the perturbation caused by ๐(๐ฅ, ๐ก). The following difference scheme can be obtained for solving (33):
๓ตจ๓ตจ ๓ตจ๓ตจ ๓ตจ๐ฟ๐ฅ ๐ข๓ตจ โ๐ขโโ โค ๓ตจ ๓ตจ1 , 2 ๓ตจ๓ตจ ๓ตจ๓ตจ ๓ตจ๐ฟ๐ฅ ๐ข๓ตจ โ๐ขโ โค ๓ตจ ๓ตจ1 . โ6
A๐ฟ๐ก๐ผ V๐๐ = ๐ท (๐ฟ๐ฅ2 V๐๐ + ๐ฟ๐ฅ2 V๐๐โ๐ ) + A๐ (2V๐๐โ1 โ V๐๐โ2 , V๐๐โ๐ , ๐ฅ๐ , ๐ก๐ ) ,
(29)
V๐๐ = ๐ (๐ฅ๐ , ๐ก๐ ) + ๐๐๐ ,
The following lemma will be used in the proof of the stability and convergence analysis. ๐
Lemma 6 (see [15]). Assume that {๐น | ๐ โฅ 0} is a nonnegative sequence and satisfies
0 โค ๐ โค ๐, โ๐ โค ๐ โค 0,
(34) (35)
V0๐ = ๐พ (๐ก๐ ) , ๐ = ๐ฝ (๐ก๐ ) , V๐
(36) 1 โค ๐ โค ๐.
Denote ๐
๐น๐+1 โค ๐ด + ๐ต๐โ๐น๐ ,
๐ = 0, 1, . . . ;
๐=1
then, ๐น๐+1 โค ๐ด exp (๐ต๐๐) , ๐ = 0, 1, 2, . . . ,
๐๐๐ = V๐๐ โ ๐ข๐๐ ,
(30)
(31)
0 โค ๐ โค ๐, โ๐ โค ๐ โค ๐.
(37)
Definition 8. Assume that ๐ข๐๐ satisfies (24)โ(26) and V๐๐ satisfies (34)โ(36); then, a numerical scheme for (3)โ(5) is stable if one has ๓ตฉ๓ตฉ ๐ ๓ตฉ๓ตฉ ๓ตจ ๓ตจ ๓ตฉ๓ตฉ๐ ๓ตฉ๓ตฉ โค ๐ถ max ๓ตจ๓ตจ๓ตจ๐๐ ๓ตจ๓ตจ๓ตจ , (38) ๓ตฉ ๓ตฉโ โ๐โค๐โค0 ๓ตจ ๓ตจ1 where ๐ถ is a bounded constant independent of โ and ๐.
where ๐ด and ๐ต are nonnegative constants. Theorem 7. The difference scheme (24)โ(26) has a unique solution. ๐ Proof. Denote ๐ข๐ = [๐ข1๐ , ๐ข2๐ , . . . , ๐ข๐โ1 ]๐ ; the difference scheme (24)โ(26) is a linear tridiagonal system M๐ข๐ = ๐, ๐ and is where ๐ only depends on ๐ข๐ (๐ โค ๐ โ 1), ๐ข0๐ , and ๐ข๐ ๐ independent of ๐ข .
1 ๐ท 10 2๐ท 1 ๐ท M = tridiag ( , โ , + โ ), 12๐ โ2 12๐ โ2 12๐ โ2
(32)
where ๐ = ๐๐ผ ฮ(2 โ ๐ผ). We can see that M is a strictly diagonally dominant coefficient matrix. Thus, scheme (24)โ(26) has a unique solution.
Theorem 9. Assume ๐ข(๐ฅ, ๐ก) โ ๐ถ6,2 ([0, 1] ร (โ๐ , ๐]) is the solution of (3)โ(5); the difference scheme (24)โ(26) is stable with respect to the initial perturbation of ๐๐๐ ; that is, โ๐๐ โโ โค ฬ maxโ๐โค๐โค0 |๐๐ |1 , where ๐ถ ฬ is a positive constant independent ๐ถ of โ and ๐. Proof. Subtracting (34)โ(36) from (24)โ(26), respectively, we can obtain the following equations: A๐ฟ๐ก๐ผ ๐๐๐ = ๐ท (๐ฟ๐ฅ2 ๐๐๐ + ๐ฟ๐ฅ2 ๐๐๐โ๐ ) + A๐๐๐ ,
(39)
๐๐๐ = ๐๐๐ , 0 โค ๐ โค ๐, โ๐ โค ๐ โค 0,
(40)
๐0๐ = 0, ๐ = 0, ๐๐
(41) 1 โค ๐ โค ๐,
To discuss the stability of the difference scheme (24)โ(26), we consider the following problem: A
๐๐ผ V ๐2 V (๐ฅ, ๐ก) ๐2 V (๐ฅ, ๐ก โ ๐ ) = ๐ท ( + ) ๐๐ก๐ผ ๐๐ฅ2 ๐๐ฅ2
where ๐๐๐ = ๐(2V๐๐โ1 โ V๐๐โ2 , V๐๐โ๐ , ๐ฅ๐ , ๐ก๐ ) โ ๐(2๐ข๐๐โ1 โ ๐ข๐๐โ2 , ๐ข๐๐โ๐ , ๐ฅ๐ , ๐ก๐ ). Multiplying โ๐ฟ๐ก๐ผ ๐๐๐ on both sides of (39) and summing up for ๐ from 1 to ๐ โ 1, we obtain ๐โ1
โ โ (A๐ฟ๐ก๐ผ ๐๐๐ ) (๐ฟ๐ก๐ผ ๐๐๐ )
+ A๐ (V (๐ฅ, ๐ก) , V (๐ฅ, ๐ก โ ๐ ) , ๐ฅ, ๐ก) ,
๐=1
(๐ฅ, ๐ก) โ (0, 1) ร (0, ๐] , V (๐ฅ, ๐ก) = ๐ (๐ฅ, ๐ก) + ๐ (๐ฅ, ๐ก) , ๐ฅ โ [0, 1] , ๐ก โ [โ๐ , 0] , V (0, ๐ก) = ๐พ (๐ก) ,
๐โ1
(33)
= ๐ทโ โ (๐ฟ๐ฅ2 ๐๐๐ + ๐ฟ๐ฅ2 ๐๐๐โ๐ ) (๐ฟ๐ก๐ผ ๐๐๐ ) ๐=1
๐โ1
+ โ โ (A๐๐๐ ) (๐ฟ๐ก๐ผ ๐๐๐ ) ,
๐ข (1, ๐ก) = ๐ฝ (๐ก) ,
๐=1
๐ก โ (0, ๐] ,
1 โค ๐ โค ๐ โ 1, 0 โค ๐ โค ๐ โ 1.
(42)
Journal of Function Spaces
5 ๐โ1
Then, each term of (42) will be estimated. From the discrete Green formula and inequality โ(๐+๐)2 โฅ โ2(๐2 +๐2 ), we have
๐โ๐ ๐ ) [๐ฟ๐ฅ ๐๐+1/2 โ
โ โ (๐ฟ๐ฅ ๐๐+1/2 ๐=0 [
๐โ1
โ โ (A๐ฟ๐ก๐ผ ๐๐๐ ) (๐ฟ๐ก๐ผ ๐๐๐ )
๐โ1
๓ตฉ ๓ตฉ2 โ2 ๐โ1 = ๓ตฉ๓ตฉ๓ตฉ๓ตฉ๐ฟ๐ก๐ผ ๐๐ ๓ตฉ๓ตฉ๓ตฉ๓ตฉ + โ โ (๐ฟ๐ฅ2 (๐ฟ๐ก๐ผ ๐๐๐ )) (๐ฟ๐ก๐ผ ๐๐๐ ) 12 ๐=1
(43)
โค
๓ตฉ2 โ2 ๓ตฉ๓ตฉ ๐ผ ๐ ๓ตฉ๓ตฉ2 2 ๓ตฉ๓ตฉ ๐ผ ๐ ๓ตฉ๓ตฉ2 ๓ตฉ ๓ตฉ๓ตฉ๐ฟ ๐ฟ ๐ ๓ตฉ๓ตฉ โฅ ๓ตฉ๓ตฉ๐ฟ๐ก ๐ ๓ตฉ๓ตฉ . = ๓ตฉ๓ตฉ๓ตฉ๓ตฉ๐ฟ๐ก๐ผ ๐๐ ๓ตฉ๓ตฉ๓ตฉ๓ตฉ โ ๓ตฉ 12 ๓ตฉ ๐ฅ ๐ก ๓ตฉ 3๓ตฉ
๓ตจ2 ๓ตจ ๓ตจ2 ๓ตจ๓ตจ ๓ตจ๓ตจ๐ฟ๐ฅ ๐๐โ๐ ๓ตจ๓ตจ๓ตจ + ๓ตจ๓ตจ๓ตจ๐ฟ๐ฅ ๐๐ ๓ตจ๓ตจ๓ตจ ๓ตจ ๓ตจ1 ๓ตจ ๓ตจ 1 + โ (๐๐โ๐โ1 โ ๐๐โ๐ ) 2 ๐=1
From the discrete Green formula and inequality โ๐๐ โฅ โ(๐ + ๐2 )/2, we have ๐ทโ โ ๐=1
๓ตจ๓ตจ ๐ ๓ตจ๓ตจ2 ๓ตจ๓ตจ๓ตจ๐ฟ ๐๐โ๐ ๓ตจ๓ตจ๓ตจ2 + ๓ตจ๓ตจ๓ตจ๐ฟ ๐0 ๓ตจ๓ตจ๓ตจ2 } ๓ตจ ๐ฅ ๓ตจ๓ตจ1 ๓ตจ๓ตจ ๐ฅ ๓ตจ๓ตจ1 ๐ท { ๓ตจ๓ตจ๓ตจ๐ฟ๐ฅ ๐ ๓ตจ๓ตจ๓ตจ1 + ๐๐โ1 ๓ตจ = } ๐{ 4 2 } {
(๐ฟ๐ฅ2 ๐๐๐ ) (๐ฟ๐ก๐ผ ๐๐๐ ) ๐โ1
๐ ๐ = โ๐ทโ โ (๐ฟ๐ฅ ๐๐+1/2 ) (๐ฟ๐ก๐ผ ๐ฟ๐ฅ ๐๐+1/2 )=โ ๐=0
๐ท ๐
+
๐โ1
๐ ๐ โ
โ โ (๐ฟ๐ฅ ๐๐+1/2 ) [๐ฟ๐ฅ ๐๐+1/2 ๐=0 [ ๐โ1
๐ท {๓ตจ๓ตจ ๓ตจ2 ๓ตจ๓ตจ๐ฟ๐ฅ ๐๐ ๓ตจ๓ตจ๓ตจ { ๓ตจ ๓ตจ1 ๐ {
From the Cauchy-Schwarz inequality, Lemma 4, and (7), we have ๐โ1
โ โ (A๐๐๐ ) (๐ฟ๐ก๐ผ ๐๐๐ ) ๐=1
โ โ (๐๐โ๐โ1 โ ๐๐โ๐ ) โจ๐ฟ๐ฅ ๐๐ , ๐ฟ๐ฅ ๐๐ โฉ
โค
๐=1
0
โค
๓ตจ๓ตจ ๓ตจ2 ๓ตจ ๓ตจ2 ๓ตจ๓ตจ ๐ ๓ตจ๓ตจ2 ๓ตจ๓ตจ๐ฟ๐ฅ ๐๐ ๓ตจ๓ตจ๓ตจ + ๓ตจ๓ตจ๓ตจ๐ฟ๐ฅ ๐0 ๓ตจ๓ตจ๓ตจ } ๐ท { ๓ตจ๓ตจ๓ตจ๐ฟ๐ฅ ๐ ๓ตจ๓ตจ๓ตจ1 ๓ตจ ๓ตจ ๓ตจ ๓ตจ 1 1 โ ๐๐โ1 }=โ๐ { 2 2 } { 2 ๓ตจ ๓ตจ๓ตจ ๓ตจ2 ๓ตจ ๓ตจ๓ตจ๐ฟ ๐๐ ๓ตจ๓ตจ ๐โ1 ๓ตจ๓ตจ๐ฟ๐ฅ ๐0 ๓ตจ๓ตจ๓ตจ } ๓ตจ๓ตจ ๐ฅ ๓ตจ๓ตจ1 ๓ตจ ๓ตจ1 , โ โ (๐๐โ๐โ1 โ ๐๐โ๐ ) โ ๐๐โ1 } 2 2 ๐=1 }
โค
๐โ1 2 1 ๓ตฉ๓ตฉ ๐ผ ๐ ๓ตฉ๓ตฉ2 ๓ตฉ๓ตฉ๐ฟ๐ก ๐ ๓ตฉ๓ตฉ + ๐๐02 โ โ (2๐๐๐โ1 โ ๐๐๐โ2 ) ๓ตฉ ๓ตฉ 2๐ ๐=1 ๐โ1
+ ๐๐12 โ โ (๐๐๐โ๐ )
๐โ1
2
๐=1
๐=1
๐=0
1 ๓ตฉ๓ตฉ ๐ผ ๐ ๓ตฉ๓ตฉ2 ๓ตฉ๓ตฉ๐ฟ ๐ ๓ตฉ๓ตฉ 2๐ ๓ตฉ ๐ก ๓ตฉ ๐ ๐โ1 ๓ตจ ๓ตจ ๓ตจ ๓ตจ 2 + โ โ (๐0 ๓ตจ๓ตจ๓ตจ๓ตจ2๐๐๐โ1 โ ๐๐๐โ2 ๓ตจ๓ตจ๓ตจ๓ตจ + ๐1 ๓ตจ๓ตจ๓ตจ๓ตจ๐๐๐โ๐ ๓ตจ๓ตจ๓ตจ๓ตจ) 2 ๐=1
๐ทโ โ (๐ฟ๐ฅ2 ๐๐๐โ๐ ) (๐ฟ๐ก๐ผ ๐๐๐ ) ๐โ๐ ๐ = โ๐ทโ โ (๐ฟ๐ฅ ๐๐+1/2 ) (๐ฟ๐ก๐ผ ๐ฟ๐ฅ ๐๐+1/2 )=โ
1 ๓ตฉ๓ตฉ ๐ผ ๐ ๓ตฉ๓ตฉ2 ๓ตฉ๓ตฉ๐ฟ ๐ ๓ตฉ๓ตฉ 2๐ ๓ตฉ ๐ก ๓ตฉ ๐ ๐โ1 ๓ตจ ๓ตจ ๓ตจ ๓ตจ 2 + โ โ (A [๐0 ๓ตจ๓ตจ๓ตจ๓ตจ2๐๐๐โ1 โ ๐๐๐โ2 ๓ตจ๓ตจ๓ตจ๓ตจ + ๐1 ๓ตจ๓ตจ๓ตจ๓ตจ๐๐๐โ๐ ๓ตจ๓ตจ๓ตจ๓ตจ]) 2 ๐=1
} ๐ท {๓ตจ ๓ตจ2 โ ๐๐โ1 โจ๐ฟ๐ฅ ๐ , ๐ฟ๐ฅ ๐ โฉ} โค โ {๓ตจ๓ตจ๓ตจ๓ตจ๐ฟ๐ฅ ๐๐ ๓ตจ๓ตจ๓ตจ๓ตจ1 ๐ } { ๓ตจ๓ตจ๓ตจ๐ฟ ๐๐ ๓ตจ๓ตจ๓ตจ2 + ๓ตจ๓ตจ๓ตจ๐ฟ ๐๐ ๓ตจ๓ตจ๓ตจ2 ๐โ1 ๓ตจ ๐ฅ ๓ตจ๓ตจ1 ๓ตจ๓ตจ ๐ฅ ๓ตจ๓ตจ1 โ โ (๐๐โ๐โ1 โ ๐๐โ๐ ) ๓ตจ 2 ๐=1
๐โ1
2
๓ตจ2 ๓ตจ๓ตจ ๓ตจ๓ตจ๐ฟ๐ฅ ๐๐ ๓ตจ๓ตจ๓ตจ ๓ตจ1 ๓ตจ + โ (๐๐โ๐โ1 โ ๐๐โ๐ ) 2 ๐=1 ๐โ1
(44)
๐โ1
๐
๓ตจ2 ๓ตจ 3 ๓ตจ๓ตจ๓ตจ๓ตจ๐ฟ๐ฅ ๐๐โ๐ ๓ตจ๓ตจ๓ตจ๓ตจ1
๓ตจ๓ตจ ๓ตจ2 ๓ตจ๓ตจ๐ฟ๐ฅ ๐0 ๓ตจ๓ตจ๓ตจ } ๓ตจ ๓ตจ1 . + ๐๐โ1 2 } }
๐
0 ] โ โ (๐๐โ๐โ1 โ ๐๐โ๐ ) ๐ฟ๐ฅ ๐๐+1/2 โ ๐๐โ1 ๐ฟ๐ฅ ๐๐+1/2 ๐=1 ]
=โ
๐ท { 1 ๓ตจ๓ตจ ๓ตจ2 ๓ตจ ๓ตจ2 ๓ตจ๓ตจ๐ฟ ๐๐ ๓ตจ๓ตจ๓ตจ + ๓ตจ๓ตจ๓ตจ๐ฟ ๐๐โ๐ ๓ตจ๓ตจ๓ตจ ๓ตจ1 ๐ { 4 ๓ตจ ๐ฅ ๓ตจ1 ๓ตจ ๐ฅ { ๐โ1
2
๐โ1
๐
0 ] โ โ (๐๐โ๐โ1 โ ๐๐โ๐ ) ๐ฟ๐ฅ ๐๐+1/2 โ ๐๐โ1 ๐ฟ๐ฅ ๐๐+1/2 ๐=1 ]
๐=1
โค ๐ท ๐
1 ๓ตฉ๓ตฉ ๐ผ ๐ ๓ตฉ๓ตฉ2 ๓ตฉ ๓ตฉ2 ๓ตฉ ๓ตฉ2 ๓ตฉ๓ตฉ๐ฟ ๐ ๓ตฉ๓ตฉ + 8๐๐02 (๓ตฉ๓ตฉ๓ตฉ๐๐โ1 ๓ตฉ๓ตฉ๓ตฉ + ๓ตฉ๓ตฉ๓ตฉ๐๐โ2 ๓ตฉ๓ตฉ๓ตฉ ) ๓ตฉ ๓ตฉ ๓ตฉ ๓ตฉ 2๐ ๓ตฉ ๐ก ๓ตฉ ๓ตฉ ๓ตฉ2 + ๐๐12 ๓ตฉ๓ตฉ๓ตฉ๓ตฉ๐๐โ๐ ๓ตฉ๓ตฉ๓ตฉ๓ตฉ .
(45)
6
Journal of Function Spaces
Substituting (43)โ(45) into (42) and taking ๐ = 3/4 in (45), we obtain
From Lemma 5, we have ๓ตฉ๓ตฉ ๐ ๓ตฉ๓ตฉ ฬ max ๓ตจ๓ตจ๓ตจ๓ตจ๐๐ ๓ตจ๓ตจ๓ตจ๓ตจ . ๓ตฉ๓ตฉ๐ ๓ตฉ๓ตฉ โค ๐ถ ๓ตฉ ๓ตฉโ โ๐โค๐โค0 ๓ตจ ๓ตจ1
๐ท ๓ตจ๓ตจ ๓ตจ2 3๐ท ๓ตจ๓ตจ ๓ตจ2 ๓ตจ๓ตจ๐ฟ๐ฅ ๐๐ ๓ตจ๓ตจ๓ตจ โค ๓ตจ๓ตจ๐ฟ๐ฅ ๐๐โ๐ ๓ตจ๓ตจ๓ตจ ๓ตจ ๓ตจ ๓ตจ ๓ตจ1 1 4๐ 2๐ +
๐ท ๐โ1 ๓ตจ2 ๓ตจ โ ๐๐โ๐ ) ๓ตจ๓ตจ๓ตจ๓ตจ๐ฟ๐ฅ ๐๐ ๓ตจ๓ตจ๓ตจ๓ตจ1 โ (๐ ๐ ๐=1 ๐โ๐โ1
The proof is completed.
๐ท ๓ตจ ๓ตจ2 ๐ ๓ตจ๓ตจ๓ตจ๐ฟ ๐0 ๓ตจ๓ตจ๓ตจ ๐ ๐โ1 ๓ตจ ๐ฅ ๓ตจ1 ๓ตฉ2 ๓ตฉ ๓ตฉ2 ๓ตฉ + 6๐02 (๓ตฉ๓ตฉ๓ตฉ๓ตฉ๐๐โ1 ๓ตฉ๓ตฉ๓ตฉ๓ตฉ + ๓ตฉ๓ตฉ๓ตฉ๓ตฉ๐๐โ2 ๓ตฉ๓ตฉ๓ตฉ๓ตฉ )
Denote ๐๐๐ = ๐๐๐ โ ๐ข๐๐ , 0 โค ๐ โค ๐, โ๐ โค ๐ โค ๐; by subtracting (24)โ(26) from (19), (22), and (23), respectively, the following error equations can be obtained:
+
3 ๓ตฉ ๓ตฉ2 + ๐12 ๓ตฉ๓ตฉ๓ตฉ๓ตฉ๐๐โ๐ ๓ตฉ๓ตฉ๓ตฉ๓ตฉ 4 โค
(46)
๐ท ๓ตจ2 ๓ตจ โ ๐๐โ๐ ) ๓ตจ๓ตจ๓ตจ๓ตจ๐ฟ๐ฅ ๐๐ ๓ตจ๓ตจ๓ตจ๓ตจ1 โ (๐ ๐ ๐=1 ๐โ๐โ1
๐0๐ = 0,
1 โค ๐ โค ๐, where ๐๐๐ = ๐(2๐๐๐โ1 โ ๐๐๐โ2 , ๐๐๐โ๐ , ๐ฅ๐ , ๐ก๐ ) โ ๐(2๐ข๐๐โ1 โ ๐ข๐๐โ2 , ๐ข๐๐โ๐ , ๐ฅ๐ , ๐ก๐ ). Similar to the proof of Theorem 9, the following convergence result can be obtained.
๐โ๐ ๓ตจ2
๓ตจ๓ตจ , ๓ตจ๓ตจ1
where Lemma 5 has been used. Multiplying (46) by 4๐/๐ท, we have
Theorem 10. Assume ๐ข(๐ฅ, ๐ก) โ ๐ถ6,2 ([0, 1] ร (โ๐ , ๐]) is the solution of (3)โ(5) and {๐ข๐๐ | 0 โค ๐ โค ๐, โ๐ โค ๐ โค ๐} is the solution of (24)โ(26). Then, one has
๐โ1 ๓ตจ ๓ตจ ๓ตจ2 ๓ตจ2 ๓ตจ2 ๓ตจ๓ตจ ๓ตจ๓ตจ๐ฟ๐ฅ ๐๐ ๓ตจ๓ตจ๓ตจ โค 4 โ (๐๐โ๐โ1 โ ๐๐โ๐ ) ๓ตจ๓ตจ๓ตจ๐ฟ๐ฅ ๐๐ ๓ตจ๓ตจ๓ตจ + 4 ๓ตจ๓ตจ๓ตจ๐ฟ๐ฅ ๐0 ๓ตจ๓ตจ๓ตจ ๓ตจ ๓ตจ ๓ตจ1 ๓ตจ1 ๓ตจ1 ๓ตจ ๐=1
+
4๐ 2 ๓ตจ๓ตจ ๓ตจ2 ๓ตจ ๓ตจ2 ๐0 (๓ตจ๓ตจ๓ตจ๐ฟ๐ฅ ๐๐โ1 ๓ตจ๓ตจ๓ตจ๓ตจ1 + ๓ตจ๓ตจ๓ตจ๓ตจ๐ฟ๐ฅ ๐๐โ2 ๓ตจ๓ตจ๓ตจ๓ตจ1 ) ๐ท
+ (6 +
(52)
๐ = 0, ๐๐
๐ท ๓ตจ ๓ตจ2 ๐ ๓ตจ๓ตจ๓ตจ๐ฟ ๐0 ๓ตจ๓ตจ๓ตจ ๐ ๐โ1 ๓ตจ ๐ฅ ๓ตจ1 ๓ตจ2 ๓ตจ ๓ตจ2 ๓ตจ + ๐02 (๓ตจ๓ตจ๓ตจ๓ตจ๐ฟ๐ฅ ๐๐โ1 ๓ตจ๓ตจ๓ตจ๓ตจ1 + ๓ตจ๓ตจ๓ตจ๓ตจ๐ฟ๐ฅ ๐๐โ2 ๓ตจ๓ตจ๓ตจ๓ตจ1 ) +
3๐ท 1 2 ๓ตจ๓ตจ + ๐ ) ๓ตจ๓ตจ๐ฟ ๐ 2๐ 8 1 ๓ตจ ๐ฅ
A๐ฟ๐ก๐ผ ๐๐๐ = ๐ท (๐ฟ๐ฅ2 ๐๐๐ + ๐ฟ๐ฅ2 ๐๐๐โ๐ ) + A๐๐๐ + ๐
๐๐ , ๐๐๐ = 0, 0 โค ๐ โค ๐, โ๐ โค ๐ โค 0,
๐โ1
+(
(51)
(47)
๓ตฉ๓ตฉ ๐ ๓ตฉ๓ตฉ ๓ตฉ๓ตฉ๐ ๓ตฉ๓ตฉ โค ๐ถ (๐2โ๐ผ + โ4 ) , ๓ตฉ ๓ตฉโ
1 โค ๐ โค ๐,
(53)
where ๐ถ is a positive constant independent of โ and ๐.
๐ 2 ๓ตจ๓ตจ ๓ตจ2 ๐ ) ๓ตจ๓ตจ๐ฟ ๐๐โ๐ ๓ตจ๓ตจ๓ตจ๓ตจ1 . 2๐ท 1 ๓ตจ ๐ฅ
4. Numerical Test
Denote ๐ถ๐ = max {
4ฮ (2 โ ๐ผ) 2 ฮ (2 โ ๐ผ) 2 ๐0 , ๐1 + 6} > 0. ๐ท 2๐ท
(48)
Noticing ๐ = ๐๐ผ ฮ(2 โ ๐ผ) for 0 < ๐ < 1, we have ๐๐ผ < 1. Then, from (47), we obtain ๐โ1
๓ตจ๓ตจ ๓ตจ2 ๓ตจ2 ๓ตจ2 ๓ตจ ๓ตจ ๓ตจ๓ตจ๐ฟ๐ฅ ๐๐ ๓ตจ๓ตจ๓ตจ โค 4 โ (๐๐โ๐โ1 โ ๐๐โ๐ ) ๓ตจ๓ตจ๓ตจ๐ฟ๐ฅ ๐๐ ๓ตจ๓ตจ๓ตจ + 4 ๓ตจ๓ตจ๓ตจ๐ฟ๐ฅ ๐0 ๓ตจ๓ตจ๓ตจ ๓ตจ ๓ตจ1 ๓ตจ1 ๓ตจ1 ๓ตจ ๓ตจ ๐=1
(49)
๓ตจ2 ๓ตจ ๓ตจ2 ๓ตจ ๓ตจ2 ๓ตจ + ๐ถ๐ (๓ตจ๓ตจ๓ตจ๓ตจ๐ฟ๐ฅ ๐๐โ1 ๓ตจ๓ตจ๓ตจ๓ตจ1 + ๓ตจ๓ตจ๓ตจ๓ตจ๐ฟ๐ฅ ๐๐โ2 ๓ตจ๓ตจ๓ตจ๓ตจ1 + ๓ตจ๓ตจ๓ตจ๓ตจ๐ฟ๐ฅ ๐๐โ๐ ๓ตจ๓ตจ๓ตจ๓ตจ1 ) . From Lemmas 2 and 6, we have ๓ตจ๓ตจ ๓ตจ2 ๓ตจ๓ตจ๐ฟ๐ฅ ๐๐ ๓ตจ๓ตจ๓ตจ ๓ตจ ๓ตจ1 ๓ตจ ๓ตจ2 โค (3๐ถ๐ + 4) exp (3๐ถ๐ + 4 (1 โ ๐๐โ1 )) max ๓ตจ๓ตจ๓ตจ๓ตจ๐ฟ๐ฅ ๐๐ ๓ตจ๓ตจ๓ตจ๓ตจ1 (50) โ๐โค๐โค0 ๓ตจ ๓ตจ2 โค (3๐ถ๐ + 4) exp (3๐ถ๐ + 4) max ๓ตจ๓ตจ๓ตจ๓ตจ๐ฟ๐ฅ ๐๐ ๓ตจ๓ตจ๓ตจ๓ตจ1 . โ๐โค๐โค0
In this section, a numerical test is used to validate the performance of scheme (24)โ(26). Denote the maximum error at all grid points as ๓ตฉ ๓ตฉ ๐ (โ, ๐) = max ๓ตฉ๓ตฉ๓ตฉ๓ตฉ๐๐ โ ๐ข๐ ๓ตฉ๓ตฉ๓ตฉ๓ตฉโ ; 1โค๐โค๐
(54)
the convergence order in time and space is defined, respectively, as Rate๐ =
log (๐ (โ, ๐1 ) /๐ (โ, ๐2 )) , log (๐1 /๐2 )
log (๐ (โ1 , ๐) /๐ (โ2 , ๐)) . Rateโ = log (โ1 /โ2 )
(55)
For Rate๐ , we require โ to be fixed and small enough, while for Rateโ , ๐ should be fixed and small enough.
Journal of Function Spaces
7 ๎ฃ = 0.2
๎ฃ = 0.4 ร10โ6
ร10โ7 5
2.5
4
2
3
1.5
2
1
1
0.5
0 1 t
1
0.5
0.5 0
0
0 1 t
x
1
0.5
0.5 0
0
๎ฃ = 0.6
x
๎ฃ = 0.8
ร10โ5 1
ร10โ5 4
0.8
3
0.6
2
0.4 0.2
1
0 1
0 1
t
1
0.5
0.5 0
0
t
x
1
0.5
0.5 0
0
x
Figure 1: Error planes for ๐ = โ = 1/100 for different ๐ผ in Example 1.
โ ๐๐ฅ (2 โ 8๐ฅ + ๐ฅ2 + 6๐ฅ3 + ๐ฅ4 )
Example 1. Consider the following problem:
โ
(๐ก2+๐ผ + (๐ก โ 0.1)2+๐ผ ) .
๐๐ผ ๐ข ๐2 ๐ข (๐ฅ, ๐ก) ๐2 ๐ข (๐ฅ, ๐ก โ 0.1) = ( + ) โ ๐ข (๐ฅ, ๐ก)2 ๐๐ก๐ผ ๐๐ฅ2 ๐๐ฅ2
(57)
+ ๐ข (๐ฅ, ๐ก โ 0.1) + ๐บ (๐ฅ, ๐ก) , (๐ฅ, ๐ก) โ (0, 1) ร (0, 1] , ๐ข (๐ฅ, ๐ก) = ๐๐ฅ ๐ฅ2 (1 โ ๐ฅ)2 ๐ก2+๐ผ ,
(56) ๐ฅ โ [0, 1] , ๐ก โ [โ0.1, 0] ,
๐ข (0, ๐ก) = 0, ๐ข (1, ๐ก) = 0, ๐ก โ (0, 1] ;
the exact solution of (56) is ๐ข(๐ฅ, ๐ก) = ๐๐ฅ ๐ฅ2 (1 โ ๐ฅ)2 ๐ก2+๐ผ , and ๐บ (๐ฅ, ๐ก) = ๐2๐ฅ ๐ฅ4 (1 โ ๐ฅ)4 ๐ก4+2๐ผ โ ๐๐ฅ ๐ฅ2 (1 โ ๐ฅ)2 ฮ (3 + ๐ผ) 2 ๐ฅ 2 โ
(๐ก โ 0.1)2+๐ผ + ๐ก ๐ ๐ฅ (1 โ ๐ฅ)2 2
From Table 1, we can see the maximum errors between the numerical solution and the exact solution in the temporal directions for ๐ผ = 0.3, 0.5, 0.8, respectively, where the spatial step is fixed to be โ = 1/400. The results show that the temporal convergence order matches well the theoretical convergence order of 2 โ ๐ผ. Table 2 shows the maximum errors in the spatial directions for ๐ผ = 0.2 when the temporal step is fixed at ๐ = 1/2000. From the results, we can see that the spatial convergence order is 4, which coincides with the theoretical result. Figure 1 gives the error plane for ๐ผ = 0.2, 0.4, 0.6, 0.8, respectively. From this figure, we can see that the error becomes larger when a larger ๐ผ is taken.
5. Conclusion This paper presents a compact finite difference scheme for solving the fractional neutral parabolic differential equation with proportional delay. The unconditional stability and the
8
Journal of Function Spaces Table 1: Maximum errors and convergence order in temporal direction with โ = 1/400 and ๐ผ = 0.3, 0.5, 0.8 for Example 1.
๐ 1/100 1/200 1/300 1/400
๐ผ = 0.3 ๐(โ, ๐) 1.083๐ โ 006 3.234๐ โ 007 1.598๐ โ 007 69.694๐ โ 008
Rate๐ โ 1.744 1.739 1.737
๐ผ = 0.5 ๐(โ, ๐) 4.502๐ โ 006 1.571๐ โ 006 8.501๐ โ 007 5.501๐ โ 007
Table 2: Maximum errors and convergence order in spatial direction with ๐ = 1/2000 and ๐ผ = 0.2 for Example 1. โ 1/8 1/12 1/16 1/20
๐(โ, ๐) 7.949๐ โ 005 1.578๐ โ 005 4.976๐ โ 006 2.045๐ โ 006
Rateโ โ 3.988 4.011 3.985
global convergence of the scheme in the maximum norm are proved. The convergence order of the considered scheme is ๐(๐2โ๐ผ + โ4 ). A numerical experiment is presented to support the theoretical results and validate the efficiency of the difference scheme.
[5]
[6]
[7]
[8]
Conflicts of Interest The authors declare that there are no conflicts of interest regarding the publication of this paper.
Acknowledgments This research work is supported by the National Natural Science Foundation of China (11401591) and the Humanities and Social Science Foundation of the Ministry of Education of China (17YJC630236). The first author acknowledges the Project of the Scientific Research Foundation for the Returned Overseas Chinese Scholars, State Education Ministry (2013693).
References [1] A. V. Rezounenko and J. Wu, โA non-local PDE model for population dynamics with state-selective delay: local theory and global attractors,โ Journal of Computational and Applied Mathematics, vol. 190, no. 1-2, pp. 99โ113, 2006. [2] B. Zubik-Kowal, โSolutions for the cell cycle in cell lines derived from human tumors,โ Computational and Mathematical Methods in Medicine. An Interdisciplinary Journal of Mathematical, Theoretical and Clinical Aspects of Medicine, vol. 7, no. 4, pp. 215โ 228, 2006. [3] D. Li, C. Zhang, and W. Wang, โLong time behavior of nonFickian delay reaction-diffusion equations,โ Nonlinear Analysis: Real World Applications, vol. 13, no. 3, pp. 1401โ1415, 2012. [4] Z.-X. Sun and Z. B. Zhang, โA linearized compact difference scheme for a class of nonlinear delay partial differential equations,โ Applied Mathematical Modelling: Simulation and
[9]
[10]
[11]
[12]
[13]
[14]
[15]
Rate๐ โ 1.519 1.515 1.513
๐ผ = 0.8 ๐(โ, ๐) 3.800๐ โ 005 1.651๐ โ 005 1.014๐ โ 005 7.176๐ โ 006
Rate๐ โ 1.203 1.202 1.202
Computation for Engineering and Environmental Systems, vol. 37, no. 3, pp. 742โ752, 2013. Q. Zhang and C. Zhang, โA compact difference scheme combined with extrapolation techniques for solving a class of neutral delay parabolic differential equations,โ Applied Mathematics Letters, vol. 26, no. 2, pp. 306โ312, 2013. W. Gu, โA compact difference scheme for a class of variable coefficient quasilinear parabolic equations with delay,โ Abstract and Applied Analysis, Article ID 810352, Art. ID 810352, 8 pages, 2014. D. Li, C. Zhang, and J. Wen, โA note on compact finite difference method for reaction-diffusion equations with delay,โ Applied Mathematical Modelling: Simulation and Computation for Engineering and Environmental Systems, vol. 39, no. 5-6, pp. 1749โ1754, 2015. C. R. Jin, Z. H. Yu, and R. N. Qu, โAn implicit difference scheme for solving a neutral delay parabolic differential equation,โ Journal of Shandong University. Natural Science. Shandong Daxue Xuebao. Lixue Ban, vol. 46, no. 8, pp. 13โ16, 2011. R. Metzler and J. Klafter, โThe random walkโs guide to anomalous diffusion: a fractional dynamics approach,โ Physics Reports, vol. 339, pp. 1โ77, 2000. S. Chen, F. Liu, I. Turner, and V. Anh, โAn implicit numerical method for the two-dimensional fractional percolation equation,โ Applied Mathematics and Computation, vol. 219, no. 9, pp. 4322โ4331, 2013. F. A. Rihan, โComputational methods for delay parabolic and time-fractional partial differential equations,โ Numerical Methods for Partial Differential Equations, vol. 26, no. 6, pp. 1556โ1571, 2010. Q. Zhang, M. Ran, and D. Xu, โAnalysis of the compact difference scheme for the semilinear fractional partial differential equation with time delay,โ Applicable Analysis: An International Journal, vol. 96, no. 11, pp. 1867โ1884, 2017. Z. Sun and X. Wu, โA fully discrete difference scheme for a diffusion-wave system,โ Appl. Numer. Math, vol. 56, pp. 193โ209, 2006. S. Chen, F. Liu, P. Zhuang, and V. Anh, โFinite difference approximations for the fractional Fokker-Planck equation,โ Applied Mathematical Modelling: Simulation and Computation for Engineering and Environmental Systems, vol. 33, no. 1, pp. 256โ273, 2009. Z. Z. Sun, The numerical methods for partial equations, Science Press, Beijing, China, 2005.
Advances in
Operations Research Hindawi Publishing Corporation http://www.hindawi.com
Volume 2014
Advances in
Decision Sciences Hindawi Publishing Corporation http://www.hindawi.com
Volume 2014
Journal of
Applied Mathematics
Algebra
Hindawi Publishing Corporation http://www.hindawi.com
Hindawi Publishing Corporation http://www.hindawi.com
Volume 2014
Journal of
Probability and Statistics Volume 2014
The Scientific World Journal Hindawi Publishing Corporation http://www.hindawi.com
Hindawi Publishing Corporation http://www.hindawi.com
Volume 2014
International Journal of
Differential Equations Hindawi Publishing Corporation http://www.hindawi.com
Volume 2014
Volume 2014
Submit your manuscripts at https://www.hindawi.com International Journal of
Advances in
Combinatorics Hindawi Publishing Corporation http://www.hindawi.com
Mathematical Physics Hindawi Publishing Corporation http://www.hindawi.com
Volume 2014
Journal of
Complex Analysis Hindawi Publishing Corporation http://www.hindawi.com
Volume 2014
International Journal of Mathematics and Mathematical Sciences
Mathematical Problems in Engineering
Journal of
Mathematics Hindawi Publishing Corporation http://www.hindawi.com
Volume 2014
Hindawi Publishing Corporation http://www.hindawi.com
Volume 2014
Volume 2014
Hindawi Publishing Corporation http://www.hindawi.com
Volume 2014
#HRBQDSDฤฎ,@SGDL@SHBR
Journal of
Volume 201
Hindawi Publishing Corporation http://www.hindawi.com
Discrete Dynamics in Nature and Society
Journal of
Function Spaces Hindawi Publishing Corporation http://www.hindawi.com
Abstract and Applied Analysis
Volume 2014
Hindawi Publishing Corporation http://www.hindawi.com
Volume 2014
Hindawi Publishing Corporation http://www.hindawi.com
Volume 2014
International Journal of
Journal of
Stochastic Analysis
Optimization
Hindawi Publishing Corporation http://www.hindawi.com
Hindawi Publishing Corporation http://www.hindawi.com
Volume 2014
Volume 2014