applied mathematics and computationdownload.xuebalib.com/xuebalib.com.40610.pdfthe basis exponential...
TRANSCRIPT
Applied Mathematics and Computation 290 (2016) 111–124
Contents lists available at ScienceDirect
Applied Mathematics and Computation
journal homepage: www.elsevier.com/locate/amc
An algorithm based on exponential modified cubic B-spline
differential quadrature method for nonlinear Burgers’
equation
Mohammad Tamsir a , Vineet K. Srivastava
b , c , Ram Jiwari d , ∗
a Department of Mathematics & Statistics, DDU Gorakhpur University, Gorakhpur 273009, India b Flight Dynamics Operations Division, ISRO Telemetry, Tracking and Command Network, Bangalore 560058, India c Department of Applied Mathematics, Indian School of Mines, Dhanbad 826004, India d Department of Mathematics, Indian Institute of Technology Roorkee, Roorkee 247667, India
a r t i c l e i n f o
Keywords:
Expo-MCB-DQM
Burgers’ equation
SSP-RK54 method
Stability analysis
a b s t r a c t
In this paper, the authors developed a new differential quadrature method "exponential
modified cubic B-spline differential quadrature method (Expo-MCB-DQM)” by using expo-
nential modified cubic B-spline functions as test functions in the traditional differential
quadrature method [32]. The new method is tested on one and two dimensional nonlin-
ear Burgers’ equations. To check the efficiency and accuracy of the proposed method five
numerical problems have been considered. The numerical results of the method are com-
pared with some existing methods and found that the proposed numerical method pro-
duces more accurate results than existing methods. Stability analysis of the algorithm is
also done by using matrix stability analysis method.
© 2016 Elsevier Inc. All rights reserved.
1. Introduction
The structure of the Burgers’ equation is like a Navier–Stoke’s equation without the stress term. It is an easiest model for
understanding the various physical flows models, such as sound and shock wave theory, wave processes in thermo-elastic
medium, vorticity transportation, dispersion in porous media, mathematical modeling of turbulent fluid, hydrodynamic tur-
bulence etc. [2–6] .
In this work, the authors consider Burgers’ equations of the form
(a) One dimensional Burgers’ equation:
∂u
∂t + αu
∂u
∂x − υ
∂ 2 u
∂ x 2 = 0 ; a ≤ x ≤ b, t ∈ [0 , T ] , (1.1)
with initial conditions
u ( x, 0 ) = ψ ( x ) ; a ≤ x ≤ b (1.2)
and Dirichlet boundary conditions
u ( a, t ) = 0 = u ( b, t ) ; t ∈ [0 , T ] , (1.3)
∗ Corresponding author. Tel.: + 91 8191957249.
E-mail address: [email protected] (R. Jiwari).
http://dx.doi.org/10.1016/j.amc.2016.05.048
0 096-30 03/© 2016 Elsevier Inc. All rights reserved.
112 M. Tamsir et al. / Applied Mathematics and Computation 290 (2016) 111–124
(b) Two dimensional coupled Burger’s equations
∂u
∂t + u
∂u
∂x + v
∂u
∂y = υ
(∂ 2 u
∂ x 2 +
∂ 2 u
∂ y 2
), (1.4)
∂v ∂t
+ u
∂v ∂x
+ v ∂v ∂y
= υ
(∂ 2 v ∂ x 2
+
∂ 2 v ∂ y 2
), (1.5)
with initial conditions
u ( x, y, 0 ) = ψ 1 ( x, y ) and v ( x, y, 0 ) = ψ 2 ( x, y ) , ( x, y ) ∈ � (1.6)
Dirichlet boundary conditions:
u ( x, y, t ) = ξ ( x, y, t ) and v ( x, y, t ) = ζ ( x, y, t ) , ( x, y ) ∈ ∂�, t > 0 (1.7)
where � = { ( x, y ) : a ≤ x ≤ b, c ≤ y ≤ d } is the computational square domain and ∂� is its boundary, u ( x, t ) is the
velocity component in one dimension, u ( x, y, t ) and v ( x, y, t ) are the velocity components in two dimension; ψ , ψ 1 ,
ψ 2 , ξ and ζ are known functions; ∂u ∂t
is unsteady term, u ∂u ∂x
is the nonlinear convection term, υ( ∂ 2 u
∂ x 2 +
∂ 2 u ∂ y 2
) is the
diffusion term and υ > 0 is viscosity coefficient and α is a positive constant. Bateman [1] was introduced this type of
equations and later a steady-state solution was proposed by Burger [2, 3] .
In the last decades, an enormous work has been done on Burgers’ equation. The Burgers’ equation is solved by various
analytical and numerical schemes such as Hopf–Cole transformation [4,8] , finite difference methods [7,10,12,14,16,17,41,42] ,
finite element method [7,13,15,19] , quadratic B-spline finite elements [14] , least-squares quadratic B-spline finite element
method [11] , B-spline collocation method [18] , Quartic B-spline collocation method [39] , reproducing kernel function
method [38] , sinc differential quadrature method [22] , polynomial differential quadrature method [20,21,44–47] , Quartic
B-spline differential quadrature method [22] , modified cubic B-splines collocation method [9] , Cubic B-spline differential
quadrature methods [21,23] , modified cubic B-spline differential quadrature method [31,43] , Galerkin quadratic B-spline
finite element method [48] , Galerkin method based on modified bi-quintic B-spline functions [49] and more recently Haar
wavelets [28–30] etc.
In 1972, Bellman et al. [32] was first introduced differential quadrature method (DQM) to solve differential equations.
After that the method was improved by Quan and Chang [33,34] . In DQM, several kinds of test functions have been used to
compute the weighting coefficients such as B-spline functions , cubic B-spline functions, sinc function, Lagrange interpolation
polynomials, Legendre polynomials, quartic B-spline functions, modified cubic B-spline functions [20–22,31,43] etc.
In this paper, the authors proposed a new method “exponential modified cubic-B-spline differential quadrature method
(Expo-MCB-DQM)” and the method is applied for the numerical simulation of the Burgers’ equation. In this method, the
exponential modified cubic-B-spline basis functions are used in traditional DQM to determine the weighting coefficients.
The method transforms the Burgers’ equation into a system of the first order nonlinear ordinary differential equations. The
resulting system is solved by employing an optimal five-stage, order four strong stability-preserving time-stepping Runge–
Kutta method. The purpose of preferring SSP-RK54 scheme is to reduce storage space. The efficiency and adaptability of the
method is confirmed by taking five test problems.
2. Exponential modified cubic B-spline differential quadrature method
Let us assume that the one dimensional domain [ a , b ] is discretized into N grid points a = x 1 < x 2 · · · < x N = b uniformly
with step size x = x i +1 − x i . By the basic fundamental of DQM, the r th order spatial partial derivatives of the unknown
u ( x, t ) with respect to x are approximated at x i , i = 1 , 2 , . . . , N as follows
∂ r u ( x i , t )
∂ x r =
N ∑
j=1
a ( r )
i j u
(x j , t
), i = 1 , 2 , . . . , N, (2.1)
where a (r) i j
are the weighting coefficients of the r th order spatial partial derivatives with respect to x .
For two dimensional domain [ a, b ] × [ c, d ] , it is assumed that N and M grid points a = x 1 < x 2 · · · < x N = b and c = y 1 <
y 2 · · · < y M
= d are uniformly distributed with step size x = x i +1 − x i and y = y j+1 − y j in x and y directions, respectively.
In the same manner, the r th order spatial partial derivatives of u ( x, y, t ) with respect to x (keeping y j as fixed) and with
respect to y (keeping x i as fixed), approximated at the point ( x i , y j ) as follows
∂ r u
(x i , y j , t
)∂ x r
=
N ∑
k =1
a ( r ) ik
u
(x k , y j , t
), i = 1 , 2 , . . . , N, j = 1 , 2 , . . . , M (2.2)
∂ r u
(x i , y j , t
)∂ y r
=
M ∑
k =1
b ( r ) jk
u ( x i , y k , t ) , i = 1 , 2 , . . . , N, j = 1 , 2 , . . . , M. (2.3)
where a (r) i j
and b (r) i j
are the weighting coefficients of the r th order spatial partial derivatives with respect to x and y .
In this work, exponential cubic B-spline basis functions are used to find the weighting coefficients of one and two di-
mensional problems.
M. Tamsir et al. / Applied Mathematics and Computation 290 (2016) 111–124 113
Table 1
Coefficients of the exponential cubic B-spline functions E i and
its derivatives at the node x i .
x i −2 x i −1 x i x i +1 x i +2
E i (x ) 0 s −ph 2( phc−s )
1 s −ph 2( phc−s )
0
E ′ i (x ) 0 p(c−1)
2( phc−s ) 0 − p(c−1)
2( phc−s ) 0
E ′′ i (x ) 0 p 2 s
2( phc−s ) − p 2 s
phc−s p 2 s
2( phc−s ) 0
2.1. Exponential cubic B-spline basis functions
The exponential cubic B-spline basis functions are defined as
E i ( x ) =
1
h
3
⎧ ⎪ ⎪ ⎪ ⎪ ⎪ ⎨
⎪ ⎪ ⎪ ⎪ ⎪ ⎩
b 2 (( x i −2 − x ) − 1
p ( sinh ( p ( x i −2 − x ) ) ) ), x ∈ [ x i −2 , x i −1 )
a 1 + b 1 ( x i − x ) + c 1 exp ( p ( x i − x ) ) + d 1 exp ( −p ( x i − x ) ) , x ∈ [ x i −1 , x i )
a 1 + b 1 ( x − x i ) + c 1 exp ( p ( x − x i ) ) + d 1 exp ( −p ( x − x i ) ) , x ∈ [ x i , x i +1 )
b 2 (( x − x i +2 ) − 1
p ( sinh ( p ( x − x i +2 ) ) ) ), x ∈ [ x i +1 , x i +2 )
0 , otherwise
(2.4)
where
a 1 =
phc
phc − s , b 1 =
p
2
(c ( c − 1 ) + s 2
( phc − s ) ( 1 − c )
), c 1 =
1
4
(exp ( −ph ) ( 1 − c ) + s ( exp ( −ph ) − 1 )
( phc − s ) ( 1 − c )
)
d 1 =
1
4
(exp ( ph ) ( c − 1 ) + s ( exp ( ph ) − 1 )
( phc − s ) ( 1 − c )
), b 2 =
p
2 ( phc − s ) , c = cosh (ph ) , s = sinh (ph ) .
In Eq. (2.4) , the free parameter p is used to obtain different forms of exponential cubic B-spline functions. The set
{ E 0 , E 1 , . . . , E N , E N+1 } is chosen in such a way that it forms a basis over the domain a ≤ x ≤ b. The values of exponential
cubic B-splines and its derivatives at the nodal points are depicted in Table 1.
The basis exponential cubic B-spline basis functions are modified in such way that the resulting matrix system of equa-
tions is diagonally dominant. The exponential cubic B-spline basis functions are modified as
φ1 ( x ) = E 1 ( x ) + 2 E 0 ( x )
φ2 ( x ) = E 2 ( x ) − E 0 ( x )
φm
( x ) = E m
( x ) f or m = 3 , . . . , N − 2
φN−1 ( x ) = E N−1 ( x ) − E N+1 ( x )
φN ( x ) = E N ( x ) + 2 E N+1 ( x )
⎫ ⎪ ⎪ ⎪ ⎪ ⎬
⎪ ⎪ ⎪ ⎪ ⎭
, (2.5)
where { φ1 , φ2 , . . . , φN } forms a basis in the region a ≤ x ≤ b.
2.2. To determine the weighting coefficients
Taking r = 1 in Eq. (2.1) and substituting the values of φm
(x ) , m = 1 , 2 , . . . , N, we get a system of linear equations
φ′ m
( x i ) =
N ∑
j=1
a (1) i j
φm
(x j
), for i, m = 1 , 2 , . . . , N. (2.6)
With the help of Eq. (2.5) and Table 1, Eq. (2.6) reduces into a tri-diagonal system of equations
A
� a ( 1 ) [ i ] =
� R [ i ] , f or i = 1 , 2 , ..., M, (2.7)
where A = [ φi j ] is the coefficient matrix of order N given by:
A =
⎡
⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎣
phc−ph phc−s
s −ph 2 ( phc−s )
0 1
s −ph 2 ( phc−s )
s −ph 2 ( phc−s )
1
s −ph 2 ( phc−s )
. . . . . .
. . . s −ph
2 ( phc−s ) 1
s −ph 2 ( phc−s )
s −ph 2 ( phc−s )
1 0
s −ph 2 ( phc−s )
phc−ph phc−s
⎤
⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎦
114 M. Tamsir et al. / Applied Mathematics and Computation 290 (2016) 111–124
� a (1) [ i ] = [ a (1)
i 1 , a (1)
i 2 , . . . , a (1)
iN ] T is the weighting coefficient vector corresponding to knot point x i , and the coefficient vector
� R [ i ] = [ φ′
1 ,i , φ′ 2 ,i , . . . , φ
′ N−1 ,i , φ
′ N,i ]
T corresponding to knot point x i , i = 1 , 2 , . . . , N are evaluated as
� R [ 1 ] =
⎡
⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎣
− p ( c−1 ) phc−s
p ( c−1 ) phc−s
0
0
. . . 0
0
⎤
⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎦
, � R [ 2 ] =
⎡
⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎣
− p ( c−1 ) 2 ( phc−s )
0
p ( c−1 ) 2 ( phc−s )
0
. . . 0
0
⎤
⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎦
, · · · , � R [ N − 1 ] =
⎡
⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎣
0
0
. . . 0
− p ( c−1 ) 2 ( phc−s )
0
p ( c−1 ) 2 ( phc−s )
⎤
⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎦
, � R [ N ] =
⎡
⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎣
0
0
. . . 0
0
− p ( c−1 ) phc−s
p ( c−1 ) phc−s
⎤
⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎦
We note that the coefficient matrix A is invertible. The tri-diagonal system of equations is solved for each knot
point x i (i = 1 , 2 , . . . , N) using the Thomas algorithm, which gives the weighting coefficients a (1) i 1
, a (1) i 2
, . . . , a (1) iN−1
, a (1) iN
(i =1 , 2 , . . . , N) of the first order partial derivative.
The weighting coefficients a (2) i j
, 1 ≤ i, j ≤ N for the second order and higher order partial derivatives are determined by
the formula [35] ⎧ ⎪ ⎨
⎪ ⎩
a ( r )
i j = r
(a (
1 ) i j
a ( r−1 )
ii − a (
r−1 ) i j
x i −x j
), f or i � = j and i = 1 , 2 , 3 , . . . , N; r = 2 , 3 , . . . , N − 1
a ( r )
ii = −
N ∑
j =1 , j � = i a (
r ) i j
, for i = j, (2.8)
where a ( r−1 ) i j
and a (r) i j
are the weighting coefficients of the ( r − 1 ) th and r th order partial derivatives with respect to x .
In the same manner, the weighting coefficients b (1) i j
of the first order partial derivatives with respect to y and weighting
coefficients b (2) i j
, 1 ≤ i, j ≤ N for the second derivatives can also be computed from the formula ⎧ ⎪ ⎨
⎪ ⎩
b ( r )
i j = r
(b (
1 ) i j
b ( r−1 )
ii − b (
r−1 ) i j
x i −x j
), f or i � = j and i = 1 , 2 , 3 , . . . , N; r = 2 , 3 , . . . , N − 1
b ( r )
ii = −
N ∑
j =1 , j � = i b (
r ) i j
, for i = j,
where b ( r−1 ) i j
and b (r) i j
are the weighting coefficients of the ( r − 1 ) th and r th order partial derivatives with respect to y.
3. Exponential modified cubic B-spline differential quadrature algorithm for Burgers’ equations
First we discretize the spatial derivatives of Burgers’ Eq. (1.1) by exponential modified cubic B-spline differential quadra-
ture method then, we get the following system of nonlinear ordinary differential equation
du ( x i , t )
dt = −α u ( x i , t )
N ∑
j=1
a (1) i j
u
(x j , t
)− υ
N ∑
j=1
a (2) i j
u
(x j , t
), a ≤ x i ≤ b, t > 0 , i = 1 , 2 , . . . , N, (3.1)
Eq. (3.1) can be written as
du ( x i , t )
dt = L ( u ( x i , t ) ) , i = 1 , 2 , . . . , N. (3.2)
with the initial and boundary conditions ( 1.2 ) and ( 1.3 ).
On substituting the approximated values of the spatial derivatives by Expo-MCB-DQM, Eqs. (1.4) and (1.5) become
∂u
(x i , y j , t
)∂t
= −u
(x i , y j
) N ∑
k =1
a (1) ik
u
(x k , y j
)− v
(x i , y j
) M ∑
k =1
b (1) jk
u ( x i , y k )
+ υ
[
N ∑
k =1
a (2) ik
u
(x k , y j
)+
M ∑
k =1
b (2) jk
u ( x i , y k )
]
, (x i , y j
)∈ �, t > 0 , i = 1 , 2 , . . . , N, j = 1 , 2 , . . . , M. (3.3)
∂v (x i , y j , t
)∂t
= −u
(x i , y j
) N ∑
k =1
a (1) ik
v (x k , y j
)− v
(x i , y j
) M ∑
k =1
b (1) jk
v ( x i , y k )
+ υ
[
N ∑
k =1
a (2) ik
v (x k , y j
)+
M ∑
k =1
b (2) jk
v ( x i , y k )
]
, (x i , y j
)∈ �, t > 0 , i = 1 , 2 , . . . , N, j = 1 , 2 , . . . , M. (3.4)
M. Tamsir et al. / Applied Mathematics and Computation 290 (2016) 111–124 115
Eqs. (3.3) and ( 3.4 ) can be written as the following system of nonlinear first order ordinary differential equations:
du
(x i , y j , t
)dt
= F 1 (u
(x i , y j , t
)), i = 1 , 2 , . . . , N and j = 1 , 2 , . . . , M. (3.5)
dv (x i , y j , t
)dt
= F 2 (u
(x i , y j , t
)), i = 1 , 2 , . . . , N and j = 1 , 2 , . . . , M. (3.6)
where L, F 1 and F 2 denotes spatial nonlinear differential operator.with the initial and boundary conditions ( 1.6 ) and ( 1.7 ).
The above system of nonlinear first order ordinary differential equations with the initial conditions and boundary condi-
tions cannot be solved directly by Runge–Kutta methods. So, first we have applied boundary conditions on the systems ( 3.1 ),
( 3.5 ) and ( 3.6 ), then we have got a system of nonlinear first order ordinary differential equations with initial conditions only.
There are various methods in literature to solve a system of nonlinear first order ordinary differential equations. We pre-
ferred the optimal five-stage, order four strong stability-preserving time-stepping Runge–Kutta (SSP-RK54) method [36,37] to
solve the system of nonlinear first order ordinary differential equations. The purpose of preferring SSP-RK54 scheme is to
reduce storage space The SSP-RK54 scheme is defined through the following steps [36] :
u
(1) = u
m + 0 . 391752226571890 tL ( u
m )
u
(2) = 0 . 4 4 4370493651235 u
m + 0 . 555629506348765 u
(1) + 0 . 368410593050371 tL ( u
(1) )
u
(3) = 0 . 620101851488403 u
m + 0 . 379898148511597 u
(2) + 0 . 251891774271694 tL ( u
(2) )
u
(4) = 0 . 178079954393132 u
m + 0 . 821920045606 86 8 u
(3) + 0 . 544974750228521 tL ( u
(3) )
u
(m +1) = 0 . 517231671970585 u
(2) + 0 . 096059710526147 u
(3) + 0 . 0 636924686 6 6290 tL ( u
(3) )
+0 . 386708617503269 u
(4) + 0 . 226007483236906 tL ( u
(4) )
4. Stability analysis of the algorithm
After discretization via DQM and linearization of the non-linear term u u x , u u x + v u y and u v x + v v y by assuming u and vlocally constant [40] , Eq. (3.1) is reduced into a set of ordinary differential equations in time as
d { U } dt
= P { U } + { E } . (4.1)
Eqs. (3.3) and ( 3.4 ) are reduced into set of ordinary differential equations in time as
d { W } dt
=
[A O
O B
]{ W } + { K } , (4.2)
where,
(i) { U} = ( u 2 , u 3 , . . . , u N−1 ) is an unknown vector of the functional values at the interior grid points.
(ii) { E} is a vector containing non-homogeneous part and boundary conditions.
(iii) P = −αU i j A 1 + υA 2 .
(iv) O
′ s are null matrices.
(v) { K} = ( F , G ) T is a vector containing non-homogeneous part and boundary conditions.
(vi) { W } = ( U, V ) T , where U and V are unknown vectors of the functional values at the interior grid points:
U = ( u 22 , u 23 , . . . , u 2(M−1) , u 32 , u 33 , . . . u 3(M−1) , . . . u (M−1)2 , u (M−1)3 . . . , u (M−1)(M−1) ) ,
V = ( v 22 , v 23 , . . . , v 2(M−1) , v 32 , v 33 , . . . v 3(M−1) , . . . v (M−1)2 , v (M−1)3 , . . . , v (M−1)(N−1) ) .
(vii)
A = −U i j A 1 − V i j B 1 + υA 2 + υB 2 ,
B = −U i j A
′ 1 − V i j B
′ 1 + υA
′ 2 + υB
′ 2 ,
where A r and B r are square block diagonal matrices (N − 2) × (M − 2) of the weighting coefficients a (r) i j
, b (r) i j
(r = 1 , 2) re-
spectively as given below
A r =
⎡
⎢ ⎢ ⎢ ⎣
a (r) 22
I a (r) 23
I . . . a (r) 2(N−1)
I
a (r) 32
I a (r) 33
I . . . a (r) 3 ,N−1
I
. . . . . .
. . . . . .
a (r) (N−1)2
I a (r) (N−1)3
I . . . a (r) (N−1)(N−1)
I
⎤
⎥ ⎥ ⎥ ⎦
116 M. Tamsir et al. / Applied Mathematics and Computation 290 (2016) 111–124
Fig. 1. Eigen values of A 1 (left) and A 2 (right) for different grid sizes.
Fig. 2. Eigen values of B 1 (left) and B 2 (right) for different grid sizes.
B r =
⎡
⎢ ⎢ ⎣
M r O . . . O
O M r . . . O
. . . . . .
. . . . . .
O O . . . M r
⎤
⎥ ⎥ ⎦
; where M r =
⎡
⎢ ⎢ ⎢ ⎣
b (r) 22
b (r) 23
. . . b (r) 2(M−1)
b (r) 32
b (r) 33
. . . b (r) 3(M−1)
. . . . . .
. . . . . .
b (r) (M−1)2
b (r) (M−1)3
. . . b (r) (M−1)(M−1)
⎤
⎥ ⎥ ⎥ ⎦
I and O are the matrices of order (N − 2) × (M − 2) .
Similarly, A
′ r and B ′ r are square block diagonal matrices each of which having order (N − 2) × (M − 2) of the weighting
coefficients a (r) i j
and b (r) i j
(r = 1 , 2) respectively.
Stability of the proposed scheme for the solution of nonlinear viscous Burgers’ equation directly depends upon the sta-
bility of the system of ordinary differential Eq. (4.2) for one dimension and (4.4) for two dimensions. Stability of ( 4.2 ) and
(4.4) depends on the eigen values of the coefficient matrices P , A and B . The system ( 4.2 ) and (4.4) will be stable if the real
part of each eigen value of P and Q are either negative or zero.
The Figs. 1 and 2 show that the real part of the eigen values of the matrices A 1 , A 2 , B 1 , and B 2 are either negative or
zero for different values of grid points. The real parts of the eigen values of the matrices P , A and B are either negative or
zero since these matrices depend upon the matrices A 1 , A 2 , B 1 , and B 2 . This shows that the developed algorithm is stable.
5. Numerical experiments and discussion
In this section, five numerical problems are considered to show the accuracy and efficiency of the proposed algorithm.
The error norms L 2 and L ∞
are calculated by using the following definitions
L 2 := || u exact − u computed | | 2 =
√
h
n ∑
j=1
| u j exact − u j
computed | 2 L ∞
:= || u exact − u computed | | ∞
= max j
| u j exact − u j
computed |
⎫ ⎪ ⎬
⎪ ⎭
,
M. Tamsir et al. / Applied Mathematics and Computation 290 (2016) 111–124 117
Table 1.1
Comparison of L 2 and L ∞ errors of Expo-MCB-DQM of Problem 1 for υ = 0 . 005 with the existing methods at different
time levels.
Methods N t t = 1 . 7 t = 2 . 4 t = 3 . 1
L 2 × 10 3 L ∞ × 10 3 L 2 × 10 3 L ∞ × 10 3 L 2 × 10 3 L ∞ × 10 3
Present (p = 1) 121 0 .01 0 .00173 0 .00680 0 .0 0 0799 0 .00288 0 .0 0 0657 0 .00354
Present ( p = 0.015) 121 0 .01 0 .00173 0 .00679 0 .0 0 0799 0 .00288 0 .0 0 0657 0 .00354
MCB-DQM [31] 121 0 .01 0 .00191 0 .00777 0 .0 0 086 0 .00308 0 .0 0 065 0 .00331
QRTDQ [22] 101 0 .001 0 .109 0 .434 0 .100 0 .339 0 .091 0 .266
BS.FEM [24] 50 0 .1 0 .857 2 .576 0 .423 1 .242 0 .230 0 .680
C.S.C. [25] 50 0 .01 0 .857 2 .576 0 .423 1 .242 0 .235 0 .688
QBCM1 [26] 200 0 .01 0 .017 0 .061 0 .012 0 .058 0 .601 4 .434
QBCM2 [26] 200 0 .001 0 .358 1 .211 0 .251 0 .807 0 .630 4 .790
Galerkin [27] 200 0 .01 0 .857 2 .576 0 .423 1 .242 0 .235 0 .688
t = 2 . 5
QBCM [18] 200 0 .01 0 .0721 0 .31153 0 .0510 0 .18902
CBCM [18] 200 0 .01 2 .4664 27 .577 2 .1118 25 .1517
t = 3 . 5
MCB-CM [9] 241 0 .01 0 .0252 0 .0994 0 .0151 0 .0549 0 .0117 0 .0486
β = 0 . 5 [38] 121 0 .01 0 .38421 1 .34728 0 .49135 1 .55470 0 .525855 1 .52196
β = 1 [38] 121 0 .01 3 .08966 10 .4040 2 .72048 8 .29747 2 .12110 5 .94321
MCB-DQM [31] 121 0 .01 0 .00191 0 .00777 0 .00778 0 .00275 0 .006177 0 .04335
Present ( p = 1) 121 0 .01 0 .00173 0 .00680 0 .0 0 0729 0 .00256 0 .006152 0 .0431
Present ( p = 0.015) 121 0 .01 0 .00173 0 .00679 0 .0 0 0729 0 .00256 0 .006152 0 .0431
where u exact and u computed represent the exact and computed solutions at the node x j , respectively.
Problem 1. In this problem, the Burgers’ Eq. (1.1) with α = 1 considered over the domain [0, 1.2] with the following initial
and boundary conditions [ 19 , 21 , 31 ]
u ( x, 1 ) =
x
1 + exp
(1
4 υ
(x 2 − 1
4
)) , and u (0 , t) = 0 , u (1 . 2 , t) = 0 , for t > 1 .
The exact solution of the problem is given by
u ( x, t ) =
x t
1 +
(t t 0
)1 / 2 exp
(x 2
4 υt
) , where t 0 = exp
(1
8 υ
), for t ≥ 1 .
The numerical solutions of this example are computed with the parameter values υ = 0 . 005 , h = 0 . 01 and t = 0 . 01
at different time levels. Table 1.1 shows the comparison of the proposed algorithm in term of L 2 and L ∞
errors with the
existing methods MCB-DQM [31] , MCB-CM [9] , QBCM [18] , CBCM [18] , QRTDQ [22] , BS.FEM [24] , C.S.C. [25] , QBCM1 [26] ,
QBCM2 [26] , Galerkin [27] , β = 0 . 5 [38] , β = 1 [38] . The physical behavior of the problem for υ = 0 . 005 at different time
levels with h = 0 . 01 , t = 0 . 01 is shown in Fig. 3. The absolute errors for different time levels are also depicted in Fig. 4 . It
is observed that the absolute errors much better than those given in [31] .
Example 2. Considered the Burger’s Eq. (1.1) , for α = 1 , over the region [0, 1] with initial condition [ 20 , 31 ]
u ( x, 0 ) = sin (πx ) ,
and the boundary conditions
u ( 0 , t ) = u (1 , t) = 0 .
The exact solution of this problem is given by Cole [4] in terms of an infinite series as
u ( x, t ) =
4 πυ∑ ∞
j=1 j I j (
1 2 πυ
)sin ( jπx ) exp (− j 2 π2 υt)
I 0 (
1 2 πυ
)+ 2
∑ ∞
j=1 I j (
1 2 πυ
)cos ( jπx ) exp (− j 2 π2 υt)
,
where I j are the modified Bessel’s functions.
Table 2.1 shows the numerical solution of the problem with parameter p = 1 , υ = 1 . 0 , h = 0 . 0125 , h = 0 . 025 and t =10 −4 at t = 0 . 1 , and a comparison is made with [ 18 , 31 ]. It is found that Expo-MCB-DQM produces the results similar to
those given in [18] at the half of the grid points and better than those given in [31] . Also, the numerical solution of the
problem with parameter value p = 1 , υ = 0 . 1 , h = 0 . 025 and t = 0 . 004 at different time levels reported in Table 2.2 . It
is clear that our results are much better than obtained in [ 9 , 18 , 20 , 31 ]. The physical behavior of solutions are depicted in
Fig. 5 at υ = 0 . 1 and υ = 1 . 0 for t ≤ 1 with h = 0 . 025 and t = 0 . 0 0 01 .
118 M. Tamsir et al. / Applied Mathematics and Computation 290 (2016) 111–124
Fig. 3. Physical behavior of Expo-MCB-DQM solutions of Problem 1 for υ = 0 . 005 at different time levels with h = 0 . 01 , t = 0 . 01 .
Fig. 4. Absolute errors in the Expo-MCB-DQM numeric solutions of Problem 1 for υ = 0 . 005 at different time levels with h = 0 . 01 , t = 0 . 01 .
Table 2.1
Comparison of Expo-MCB-DQM solutions of Problem 2 for υ = 1 . 0 with the existing solutions and exact solutions at t = 1.0.
x [18] h = 0 . 0125 [18] h = 0 . 00625 MCB-DQM [31]
h = 0 . 025
Present (p = 1)
h = 0 . 025
MCB-DQM [31]
h = 0 . 0125
Present (p = 1)
h = 0 . 0125
Exact
t = 10 −5 t = 10 −5 t = 10 −4 t = 10 −4 t = 10 −4 t = 10 −4
0 .1 0 .10952 0 .10953 0 .109530 0 .109541 0 .109526 0 .109538 0 .10954
0 .2 0 .20975 0 .20977 0 .209771 0 .209795 0 .209766 0 .209792 0 .20979
0 .3 0 .29184 0 .29186 0 .291860 0 .291899 0 .291855 0 .291896 0 .29190
0 .4 0 .34785 0 .34788 0 .347874 0 .347927 0 .347869 0 .347923 0 .34792
0 .5 0 .37149 0 .37153 0 .371517 0 .371581 0 .371512 0 .371577 0 .37158
0 .6 0 .35896 0 .35900 0 .358981 0 .359049 0 .358975 0 .359045 0 .35905
0 .7 0 .30983 0 .30986 0 .309845 0 .309909 0 .309839 0 .309904 0 .30991
0 .8 0 .22776 0 .22778 0 .227773 0 .227822 0 .227766 0 .2278217 0 .22782
0 .9 0 .12065 0 .12067 0 .120666 0 .120691 0 .120659 0 .120686 0 .12069
M. Tamsir et al. / Applied Mathematics and Computation 290 (2016) 111–124 119
Table 2.2
Comparison of Expo-MCB-DQM solutions of Problem 2 for υ = 0 . 1 with the existing and exact solutions.
x t Dag et al. [18]
h = 0 . 0125
Mittal and Jain [9]
h = 0 . 025
Korkmaz [20] h = 0 . 025 Arora and Singh
[31] h = 0 . 025
Present (p = 1)
h = 0 . 025
Exact
t = 10 −4 t = 0 . 0025 t = 0 . 00125 t = 0 . 004 t = 0 . 004
0 .25 0 .4 0 .30890 0 .30892 0 .30910 0 .3089280 0 .308893 0 .30889
0 .6 0 .24075 0 .24077 0 .24093 0 .2407550 0 .240738 0 .24074
0 .8 0 .19569 0 .19572 0 .19586 0 .1956840 0 .195674 0 .19568
1 .0 0 .16258 0 .16261 0 .16274 0 .1625700 0 .162563 0 .16256
3 .0 0 .02720 0 .02718 0 .02720 0 .0272047 0 .0272005 0 .02720
0 .50 0 .4 0 .56965 0 .56970 0 .56973 0 .5696530 0 .569631 0 .56963
0 .6 0 .44723 0 .44729 0 .44736 0 .4472170 0 .447202 0 .44721
0 .8 0 .35925 0 .35930 0 .35943 0 .3592450 0 .359231 0 .35924
1 .0 0 .29192 0 .29195 0 .29213 0 .2919250 0 .291910 0 .29192
3 .0 0 .04019 0 .04016 0 .04032 0 .0402085 0 .0402023 0 .04021
0 .75 0 .4 0 .62538 0 .62520 0 .62573 0 .6253490 0 .625424 0 .62544
0 .6 0 .48715 0 .48694 0 .48760 0 .4872040 0 .487194 0 .48721
0 .8 0 .37385 0 .37365 0 .37434 0 .3739350 0 .373901 0 .37392
1 .0 0 .28741 0 .28724 0 .28788 0 .2874930 0 .287456 0 .28747
3 .0 0 .02976 0 .02974 0 .029881 0 .0297753 0 .0297697 0 .02977
Fig. 5. Physical behavior of Expo-MCB-DQM of Problem 2 at υ = 0 . 1 (left) and at υ = 1 (right) for t ≤ 1 with h = 0 . 025 and t = 0 . 0 0 01 .
Table 3.1
Comparison of L 2 and L ∞ errors of Expo-MCB-DQM of Problem 3 with the existing methods.
N Mittal and Jain [9] Kaysar et al. [46] Jiwari et al. [47] Present scheme
L ∞ L 2 L ∞ L 2 L ∞ L 2 L ∞ L 2
10 4 .62E −07 3 .28E −07 4 .881E −07 3 .455E −07 4 .708E −08 6 .459E −08 1 .467E −07 6 .330E −08
20 1 .16E −07 8 .19E −08 1 .431E −07 1 .012E −07 1 .091E −08 4 .465E −09 3 .029E −08 1 .014E −08
40 2 .907E −08 2 .047E −08 5 .668E −08 4 .003E −08 1 .980E −09 2 .786E −10 3 .956E −09 1 .207E −09
80 7 .271E −09 5 .119E −09 3 .499E −08 4 .002E −08 7 .18E −09 2 .665E −10 8 .861E −11 1 .322E −10
Problem 3. In this problem, the Burgers’ Eq. (1.1) with α = 1 is considered with the following exact solution [9,47]
u ( x, t ) = 2 πυsin (πx ) exp (−π2 υt)
σ + cos (πx ) exp (−π2 υt) , for x ∈ ( 0 , 1 ) and t ≥ 0 ,
where, the parameter σ > 1 .
The initial and boundary conditions are taken from the exact solution. The comparison of L ∞
and L 2 errors with parame-
ters υ = 0 . 005 , σ = 100 , p = 0 . 1 , t = 0 . 01 at t = 1 for different grid sizes, is reported in Table 3.1 . It is evident that present
method results are much better than the results obtained in [9,46] and are in good agreement with the results obtained in
[47] . Absolute error for parameters σ = 2 , υ = 0 . 001 at t = 0 . 5 at different t is depicted in Fig. 6.
120 M. Tamsir et al. / Applied Mathematics and Computation 290 (2016) 111–124
Fig. 6. Absolute error of Problem 3 for σ = 2 , υ = 0 . 001 at t = 0 . 5 for different time step t.
Table 4.1
Errors and rate of convergence for u component for υ = 10 −2 , t = 0.0 0 01 at t = 1 . 0 .
Grid L 2 L ∞
Srivastava et al. [42] Shukla et al. [43] Expo-MCB-DQM Srivastava et al. [42] Shukla et al. [43] Expo-MCB-DQM
p = 10 ROC p = 10 ROC
4 × 4 8 .5708E −02 1 .6388E −02 1 .5865E −02 – 9 .7046E −02 2 .8788E −03 2 .3325E −03 –
8 × 8 4 .9429E −02 1 .9286E −03 1 .8037E −03 3 .137 4 .6886E −02 1 .9572E −04 1 .6816E −04 3 .794
16 × 16 1 .9192E −02 3 .9474E −04 3 .8329E −04 2 .234 2 .0467E −02 2 .0486E −05 1 .9610E −05 3 .100
32 × 32 8 .6812E −03 8 .1181E −05 8 .0461E −05 2 .252 9 .0744E −03 2 .2202E −06 2 .1967E −06 3 .158
64 × 64 – 1 .5322E −05 1 .5355E −05 2 .387 – 2 .1838E −07 2 .1795E −07 3 .333
Problem 4. In this problem, the two dimensional Burgers’ Eqs. (1.4) and ( 1.5 ) are considered over the domain D ={ ( x, y ) : 0 ≤ x ≤ 1 , 0 ≤ y ≤ 1 } with an exact solution generated by using the Hopf–Cole transformation [8]
u ( x, y, t ) =
3
4
− 1
4
(1 + e
Re (4 y −4 x −t) 32
)
v ( x, y, t ) =
3
4
+
1
4
(1 + e
Re (4 y −4 x −t) 32
)The initial and boundary conditions are taken from the exact solutions. In this problem, numerical solutions are com-
puted with the parameters p = 10 , υ = 10 −2 , t = 0.0 0 01 at t = 1 . 0 for different grid sizes and reported in Tables 4.1 and
4.2 in the form of errors and the rate of convergence for u and v , respectively. It is found that that the Expo-MCB-DQM
performs much better than [42,43] and gives more than quadratic rate of convergence (see Figs. 7 and 8 ).
Problem 5. In this problem, the Eqs. (1.4) and ( 1.5 ) are solved in the computational domain 0 ≤ x ≤ 0 . 5 , 0 ≤ y ≤ 0 . 5 as taken
in [43] with initial conditions
u ( x, y, 0 ) = sin (πx ) + cos (πy ) , v ( x, y, 0 ) = x + y,
}; 0 ≤ x ≤ 0 . 5 , 0 ≤ y ≤ 0 . 5 ,
M. Tamsir et al. / Applied Mathematics and Computation 290 (2016) 111–124 121
Fig. 7. Comparison of (a) numerical and (b) exact solution for component u and v of Problem 4 for Re = 100 , h = 0 . 05 , t = 0 . 0 0 01 at t = 1 .
Fig. 8. Comparison of (a) numerical and (b) exact solution for component u and v of Problem 4 for Re = 200 , h = 0 . 05 , t = 0 . 0 0 01 at t = 1 .
122 M. Tamsir et al. / Applied Mathematics and Computation 290 (2016) 111–124
Table 4.2
Errors and rate of convergence for v component for υ = 10 −2 , t = 0.0 0 01 at t = 1 . 0 .
Grid L 2 L ∞
Srivastava et al. [42] Shukla et al. [43] Expo-MCB-DQM Srivastava et al . [42] Shukla et al. [43] Expo-MCB-DQM
p = 10 ROC p = 10 ROC
4 × 4 8 .5708E −02 1 .6388E −02 1 .5865E −02 – 9 .7046E −02 2 .8788E −03 2 .3325E −03 –
8 × 8 4 .9431E −02 1 .9286E −03 1 .8037E −03 3 .137 4 .6887E −02 1 .9573E −04 1 .6816E −04 3 .794
16 × 16 1 .9196E −02 3 .9474E −04 3 .8329E −04 2 .234 2 .0471E −02 2 .0486E −05 1 .9610E −04 3 .100
32 × 32 8 .6878E −03 8 .1181E −05 8 .0461E −05 2 .252 9 .0813E −03 2 .2202E −06 2 .1967E −06 3 .158
64 × 64 – 1 .5322E −05 1 .5355E −05 2 .387 – 2 .1838E −07 2 .1795E −07 3 .333
Table 5.1
Comparison of the results of Expo-MCB-DQM for Re = 50 with grid size 20 × 20 and t = 0.0 0 01 at t = 0.625.
Grid ( x, y ) u ( x, t ) v ( x, t )
I-LFDM [42] MCB-DQM [43] Present I-LFDM [42] MCB-DQM [43] Present
(0 .1, 0.1) 0 .97146 0 .97056 0 .970558 0 .09869 0 .09842 0 .098419
(0 .3, 0.1) 1 .15280 1 .15152 1 .15152 0 .14158 0 .14107 0 .141070
(0 .2, 0.2) 0 .86308 0 .86244 0 .862434 0 .16754 0 .16732 0 .167317
(0 .4, 0.2) 0 .97985 0 .98078 0 .980779 0 .17111 0 .17223 0 .172228
(0 .1, 0.3) 0 .66316 0 .66336 0 .663354 0 .26378 0 .26380 0 .263801
(0 .3, 0.3) 0 .77233 0 .77226 0 .772256 0 .22655 0 .22653 0 .226526
(0 .2, 0.4) 0 .58181 0 .58273 0 .582728 0 .32851 0 .32935 0 .329347
(0 .4, 0.4) 0 .75862 0 .76179 0 .761787 0 .32502 0 .32884 0 .328842
Fig. 9. Physical behavior the numerical solution of Problem 5 for Re = 100 , h = 0 . 05 , t = 0 . 0 0 01 at t = 1 .
and boundary conditions
u ( 0 , y, t ) = cos (πy ) , u ( 0 . 5 , y, t ) = 1 + cos (πy ) , v ( 0 , y, t ) = y, v ( 0 . 5 , y, t ) = 0 . 5 + y,
⎫ ⎪ ⎬
⎪ ⎭
; 0 ≤ y ≤ 0 . 5 , t ≥ 0 ,
u ( x, 0 , t ) = 1 + sin (πx ) , u ( x, 0 . 5 , t ) = sin (πx ) , v ( x, 0 , t ) = x,
v ( x, 0 . 5 , t ) = x + 0 . 5
⎫ ⎪ ⎬
⎪ ⎭
; 0 ≤ x ≤ 0 . 5 , t ≥ 0 .
Table 5.1 shows the comparison of numerical solution for the component u ( x, t ) and v ( x, t ) for Re = 50, grid size 20 × 20 ,
t = 0 . 0 0 01 at t = 0.625 with [ 42 , 43 ]. The table concludes that the present results are good in agreement with the results
[42,43] (see Figs. 9–11 ).
6. Conclusions
In this paper, the authors developed a new differential quadrature method “Expo-MCB-DQM” to solve nonlinear par-
tial differential equations. The proposed method is tested on well known nonlinear Burgers’ equations. Finally, the authors
summarize the outcomes of this analysis as follows:
M. Tamsir et al. / Applied Mathematics and Computation 290 (2016) 111–124 123
Fig. 10. Physical behavior the numerical solution of Problem 5 for Re = 100 , h = 0 . 05 , t = 0 . 0 0 01 at t = 2 .
Fig. 11. Physical behavior the numerical solution of Problem 5 for Re = 100 , h = 0 . 05 , t = 0 . 0 0 01 at t = 3 .
(i) A different technique based on exponential modified cubic-B-spline functions is proposed to find the weighting coef-
ficients of differential quadrature method than the traditional technique of Lagrange interpolation [32] .
(ii) To the best knowledge of the authors, this is new differential quadrature technique for solving differential equations.
(iii) The new proposed technique gives better results than the results discussed in [9,18,22,24–27,31,38,42,43,47] and good
accuracy for small number of grid points.
(iv) The present method with some modifications can be easily extended to solve model equations in two or higher
dimensional problems including mechanical, physical or biophysical effects, such as nonlinear convection, reaction,
linear diffusion and dispersion.
(v) The low memory storage, ease of the implementation and good accuracy for small number of grid points are the
advantage of the proposed method.
Acknowledgment
The authors are very thankful to the reviewers for their valuable suggestions to improve the quality of the paper .
References
[1] H. Bateman , Some recent researches on the motion of fluids, Mon. Weather Rev. 43 (1915) 163–170 .
[2] J.M. Burgers , Mathematical example illustrating relations occurring in the theory of turbulent fluid motion, Trans. R. Neth. Acad. Sci. Amst. 17 (1939)1–53 .
[3] J.M. Burgers , A mathematical model illustrating the theory of turbulence, Adv. Appl. Mech. 1 (1948) 171–199 . [4] J.D. Cole , On a quasilinear parabolic equations occurring in aerodynamics, Q. Appl. Math. 9 (1951) 225–236 .
[5] J.D. Logan , An Introduction To Nonlinear Partial Differential Equations, Wily-Interscience, New York, 1994 .
[6] S.E. Esipov , Coupled Burgers’ equations: a model of poly-dispersive sedimentation, Phys. Rev. 52 (1995) 3711–3718 . [7] C.A.J. Fletcher , A comparison of finite element and finite difference of the one and two dimensional Burgers’ equations, J. Comput. Phys. 51 (1983)
159–188 . [8] C.A.J. Fletcher , Generating exact solutions of the two-dimensional Burgers’ equation, Int. J. Numer. Methods Fluids 3 (1983) 213–216 .
[9] R.C. Mittal , R.K. Jain , Numerical solutions of nonlinear Burgers’ equation with modified cubic B-splines collocation method, Appl. Math. Comput. 218(2012) 7839–7855 .
124 M. Tamsir et al. / Applied Mathematics and Computation 290 (2016) 111–124
[10] S. Kutluay , A.R. Bahadir , A. Ozdes , Numerical solution of one-dimensional Burgers’ equation: explicit and exact-explicit finite difference methods, J.Comput. Appl. Math. 103 (1999) 251–261 .
[11] S. Kutulay , A. Esen , I. Dag , Numerical solutions of the Burgers’ equation by the least-squares quadratic B-spline finite element method, J. Comput.Appl. Math. 167 (2004) 21–33 .
[12] W. Liao , An implicit fourth-order compact finite difference scheme for one-dimensional Burgers’ equation, Appl. Math. Comput. 206 (2008) 755–764 . [13] T. Ozis , E.N. Aksan , A. Ozdes , A finite element approach for solution of Burgers’ equation, Appl. Math. Comput. 139 (2003) 417–428 .
[14] T. Ozis , A. Esen , S. Kutluay , Numerical solution of Burgers’ equation by quadratic B-spline finite elements, Appl. Math. Comput. 165 (2005) 237–249 .
[15] A. Dogan , A. Galerkin , Finite element approach to Burgers’ equation, Appl. Math. Comput. 157 (2004) 331–346 . [16] I.A . Hassanien , A .A . Salama , H.A . Hosham , Fourth-order finite difference method for solving Burgers’ equation, Appl. Math. Comput. 170 (2005)
781–800 . [17] M.K. Kadalbajoo , K.K. Sharma , A. Awasthi , A parameter-uniform implicit difference scheme for solving time-dependent Burgers’ equation, Appl. Math.
Comput. 170 (2005) 1365–1393 . [18] I. Dag , D. Irk , A. Sahin , B-Spline collocation methods for numerical solutions of the Burgers’ equation, Math. Probl. Eng. 5 (2005) 521–538 .
[19] E.N. Aksan , Quadratic B-spline finite element method for numerical solution of the Burgers’ equation, Appl. Math. Comput. 174 (2006) 884–896 . [20] A. Korkmaz , I. Dag , Shock wave simulations using sinc differential quadrature method, Eng. Comput. Int. J. Comput. Aided Eng. Softw. 28 (6) (2011)
654–674 .
[21] A. Korkmaz , I. Dag , Polynomial based differential quadrature method for numerical solution of nonlinear Burgers’ equation, J. Frankl. Inst. 348 (10)(2011) 2863–2875 .
[22] A. Korkmaz , A.M. Aksoy , I. Dag , Quartic B-spline differential quadrature method, Int. J. Nonlinear Sci. 11 (4) (2011) 403–411 . [23] A. Korkmaz , I. Dag , Cubic B-spline differential quadrature methods for the advection-diffusion equation, Int. J. Numer. Methods Heat Fluid Flow 22 (8)
(2012) 1021–1036 . [24] J. Chung , E. Kim , Y. Kim , Asymptotic agreement of moments and higher order contraction in the Burgers equation, J. Differ. Equ. 248 (10) (2010)
2417–2434 .
[25] A.H. Salas , Symbolic computation of solutions for a forced Burgers equation, Appl. Math. Comput. 216 (1) (2010) 18–26 . [26] S. Lin , C. Wang , Z. Dai , New exact traveling and non-traveling wave solutions for (2 + 1) dimensional Burgers equation, Appl. Math. Comput. 216 (10)
(2010) 3105–3110 . [27] M. Xu , R.-H. Wang , J.-H. Zhang , Q. Fang , A novel numerical scheme for solving Burgers’ equation, Appl. Math. Comput. 217 (2011) 4 473–4 482 .
[28] R. Jiwari , A hybrid numerical scheme for the numerical solution of the Burgers’ equation, Comput. Phys. Commun. 188 (2015) 59–67 . [29] R. Jiwari , Haar wavelet quasilinearization approach for numerical simulation of Burgers’ equation, Comput. Phys. Commun. 183 (2012) 2413–2423 .
[30] M. Kumar , Sapna Pandit , A composite scheme for the numerical simulation of coupled Burgers’ equation, Comput. Phys. Commun. 185 (2014) 809–817 .
[31] G. Arora , B.K. Singh , Numerical solution of Burgers’ equation with modified cubic B-spline differential quadrature method, Appl. Math. Comput. 224(1) (2013) 166–177 .
[32] R. Bellman , B.G. Kashef , J. Casti , Differential quadrature: a technique for the rapid solution of nonlinear differential equations, J. Comput. Phys. 10(1972) 40–52 .
[33] J.R. Quan , C.T. Chang , New insights in solving distributed system equations by the quadrature methods-I, Comput. Chem. Eng. 13 (1989) 779–788 . [34] J.R. Quan , C.T. Chang , New insights in solving distributed system equations by the quadrature methods-II, Comput. Chem. Eng. 13 (1989) 1017–1024 .
[35] C. Shu , Differential Quadrature and its Application in Engineering, Athenaeum Press Ltd, Great Britain, 20 0 0 .
[36] J.R. Spiteri , S.J. Ruuth , A new class of optimal high-order strong stability-preserving time-stepping schemes, SIAM J. Numer. Anal. 40 (2) (2002)469–491 .
[37] S. Gottlieb , D.I. Ketcheson , C.W. Shu , High order strong stability preserving time discretizations, J. Sci. Comput. 38 (2009) 251–289 . [38] S.-S. Xie , S. Heo , S. Kim , G. Woo , S. Yi , Numerical solution of one-dimensional Burgers’ equation using reproducing kernel function, J. Comput. Appl.
Math. 214 (2008) 417–434 . [39] B. Saka , I. Dag , Quartic B-spline collocation method to the numerical solutions of the Burgers’ equation, Chaos Solitons Fractals 32 (2007) 1125–1137 .
[40] B. Saka , A. Sahin , I. Dag , B-spline collocation algorithms for numerical solutions of the RLW equation, Numer. Methods Partial Differ. Equ. 27 (3) (2009)
581–607 . [41] V.K. Srivastava , M.K. Awasthi , M. Tamsir , A fully implicit Finite-difference solution to one-dimensional Coupled Nonlinear Burgers’ equations, Int. J.
Math. Sci. 7 (4) (2013) 23 . [42] V.K. Srivastava , M.K. Awasthi , S. Singh , An implicit logarithm finite difference technique for two dimensional coupled viscous Burgers’ equation, AIP
Adv. 3 (2013) 122105 . [43] H.S. Shukla , Mohammad Tamsir , V.K. Srivastava , J. Kumar , Numerical solution of two dimensional coupled viscous Burger equation using modified
cubic B-spline differential quadrature method, AIP Adv. 4 (2014) 117134 .
[44] R.C. Mittal , Ram Jiwari , A differential quadrature method for solving Burgers’-type equation, Int. J. Numer. Methods Heat Fluid Flow 22 (7) (2012)880–895 .
[45] R.C. Mittal , Ram Jiwari , K.K. Sharma , A numerical scheme based on differential quadrature method to solve time dependent Burgers’ equation, Eng.Comput. 30 (1) (2013) 117–131 .
[46] R.C. Mittal , Ram Jiwari , A differential quadrature method for solving Burgers’-type equation, Int. J. Numer. Methods Heat Fluid Flow 22 (7) (2012)880–895 .
[47] Ram Jiwari , R.C. Mittal , Kapil K. Sharma , A numerical scheme based on weighted average differential quadrature method for the numerical solution ofBurgers’ equation, Appl. Math. Comput. 219 (2013) 6680–6691 .
[48] S. Kutluay , Y Ucar , Numerical solutions of the coupled Burgers’ equation by the Galerkin quadratic B-spline finite element method, Math. Methods
Appl. Sci. 36 (17) (2013) 2403–2415 . [49] S. Kutluay , N.M. Yagmurlu , The modified bi-quintic B-splines for solving the two-dimensional unsteady Burgers’ equation, Eur. Int. J. Sci. Technol. 1 (2)
(2012) 23–39 .
本文献由“学霸图书馆-文献云下载”收集自网络,仅供学习交流使用。
学霸图书馆(www.xuebalib.com)是一个“整合众多图书馆数据库资源,
提供一站式文献检索和下载服务”的24 小时在线不限IP
图书馆。
图书馆致力于便利、促进学习与科研,提供最强文献下载服务。
图书馆导航:
图书馆首页 文献云下载 图书馆入口 外文数据库大全 疑难文献辅助工具