Numerical Solution of Sixth-Order Differential Equations Arising in Astrophysics by Neural Network

International Journal of Computer Applications (0975 8887)
Volume 107 - No. 6, December 2014
Numerical Solution of Sixth-Order Differential

Equations Arising in Astrophysics by Neural Network
M. Khalid
Mariam Sultana
Department of Mathematical Sciences

Federal Urdu University of Arts, Sciences & Techonology
University Road, Karachi-75300, Pakistan

Federal Urdu University of Arts, Sciences & Technology
Faheem Zaidi
Federal Urdu University of Arts, Sciences & Technology
ABSTRACT
In the current paper, a neural network method to solve sixth-order
differential equations and their boundary conditions has been presented. The idea this method incorporates is to integrate knowledge about the differential equation and its boundary conditions
into neural networks and the training sets. Neural networks are being used incessantly to solve all kinds of problems hailing a wide
range of disciplines. Several examples are given to illustrate the
efficiency and implementation of the Neural Network method.
Keywords:
Sixth Order Differential Equation, Boundary Conditions, Neural
Network
1.
INTRODUCTION
Neural networks of an artificial nature are an interesting kind of

artificial intelligence; they are unique in itself. Neural networks
mimic the learning process of the human brain in order to extract
patterns from historical data [1]. For a great number of years
this new type of expertise has been successfully employed to
a variety of real-world problems [2]. The first introduction of
perceptrons was proposed by Rosenblatt [3]. Simple perceptrons
are supervised networks; in the case that is largely unsupervised,
the network completely acclimatizes to its inputs and furthers itself
accordingly. These networks can be taught to distinguish structures
and patterns from their input. Multi-layered perceptrons (those that
have more than three layers) use multiple hidden layers.
Astrophysics gives rise to many sixth-order boundary-value
problems; the thin layers that convect are bounded by stable
layers, and the former are known to surround A-type stars. This
may be modeled by sixth-order boundary-value problems [4, 5].
Numerical analysis literature proves that very little work has been
conducted on finding the solution for sixth-order boundary-value
problems [48]. A thorough discussion has been presented regarding theorems that list conditions for the existence and uniqueness
of solutions of such problems in [9].
Looking at it from the perspective of numerals, Shen [10],

Doha and Bhrawy [1113], and Doha et al. [14] have succeeded in
constructing a competent Spectral-Galerkin algorithms that make
use of packed combinations of orthogonal polynomials to solve
elliptic equations of the second and fourth order with coefficients
that remain constant in various situations. In [1517] the respective
authors have deigned to present efficient Jacobi dual-PetrovGalerkin and Jacobi-Gauss methods of collocation to find solutions
to differential equations of some odd-order. More so, the Bernstein
polynomials have been employed for conducting a numerical
solution of differential equations of high even-order [18, 19]. Siddiqi and Twizell [7] made use of sixth-degree Splines, where the
value of Spline at the mid knots of the interpolation interval were
related through consistency relations to the corresponding values
of the even order derivatives. A B-Spline function of sixth-degree
is employed for the processing of an approximate solution for
boundary-value problems of the sixth order [20]. Furthermore,
another set of sixth-order problem related solutions are introduced,
called Septic Spline solutions [21]. Another milestone was sixthorder differential equations by Twizell and Boutayeb [5]; they
developed finite-difference methods of order two, four, six, and
eight for the solution of problems as these. Authors in [8] gave rise
to the Sinc-Galerkin method to solve sixth-order boundary-value
problems. Wazwaz [22] used decomposition and modified domain
decomposition methods for the same purpose and Bhrawy [23] put
forward a spectral Legendre-Galerkin method to solve sixth-order
boundary-value problems.
In this work, sixth order differential equations with boundary
conditions are mathematically modeled with feed-forward artificial
neural network with some adaptive parameters. The paper has been
organized as follows: Section 2 summarizes and discusses papers
that address neural network in solving Differential Equation; the
objective function creation with neural network mathematical
modeling and its training methodology are introduced in Section
3. A detailed application of the designed method along with some
discussion on the results is presented in Section 4. Comparative
analysis of the results, conclusively, is presented in Section 5.

2.
LITERATURE REVIEW
Differential equations play an imperative role in various fields

of engineering and science. The exact solution of differential
equations may not be always possible. For this reason, then,
various types of well known numerical methods such as Euler,
Runge-Kutta, Predictor-Corrector, finite element, and finite difference methods, are incorporated for solving these equations.
Although these numerical methods provide good approximations
to the solution, they may be challenging for problems on a higher
dimension. In recent years, many researchers tried to devise new
methods for solving differential equations. As such are Artificial
Neural Network (AN N ) based models are used to solve ordinary
differential equations with initial conditions.
Lee & Kang [24] initially introduced a method to solve first
order differential equation using Hopfield neural network models.
Then, another approach by Meade & Fernandez [25, 26] was
proposed for both linear and non-linear differential equations
using Splines and feed forward neural network. Artificial neural
networks based on Broyden-Fletcher-Goldfarb-Shanno (BF GS)
optimization technique for solving ordinary and partial differential
equations have been excellently presented by Lagaris et al. [27].
Furthermore, Lagaris et al. [28] investigated neural network methods for boundary value problems with irregular boundaries. Parisi
et al. [29] presented unsupervised feed forward neural networks for
the solution of differential equations. The potential of the hybrid
and optimization technique to deal with differential equation of
lower order as well as higher order has been presented by Malek
& Shekari Beidokhti [30]. Choi & Lee [31] discussed comparison
of generalizing ability on solving differential equation using back
propagation and reformulated radial basis function network. Yazdi
et al. [32] used unsupervised kernel least mean square algorithm
for solving ordinary differential equations. A new algorithm for
solving matrix Riccati differential equations has been developed
and employed by Selvaraju & Abdul Samant [33]. He et al. [34]
investigated a class of partial differential equations using multilayer neural network. Kumar & Yadav [35] surveyed multilayer
perceptrons and radial basis function neural network methods for
the solution of differential equations. Tsoulos et al. [36] solved
differential equations with neural networks using a scheme that
worked on the basis of grammatical evolution. Numerical solution
of elliptic partial differential equation using radial basis function
neural networks has been presented by Jianyu et al. [37]. Shirvany
et al. [38] proposed multilayer perceptron and radial basis function
(RBF ) neural networks with a new unsupervised training method
for numerical solution of partial differential equations. Mai-Duy
& Tran-Cong [39] discussed numerical solution of differential
equations using multiquadric radial basis function networks. Fuzzy
linguistic model in neural network to solve differential equations is
presented by Leephakpreeda [40]. Franke & Schaback [41] solved
partial differential equations by collocation using radial basis
functions. Smaoui & Al-Enezi [42] presented the dynamics of
two non-linear partial differential equations using artificial neural
networks. Differential equations with genetic programming have
been analyzed by Tsoulos & Lagaris [43]. McFall & Mahan [44]
used artificial neural network for solution of boundary value
problems with exact satisfaction of arbitrary boundary conditions.
Hoda & Nagla [45] solved mixed boundary value problems using
multilayer perceptron neural network method.
As per the review of the literatures, it reveals that authors
have taken the parameters (weights/biases) as arbitrary (random)
and the numbers of nodes in the hidden layer are considered

by a trial and error method. In this paper, a method for solving
sixth-order differential equations using feed forward neural
network is proposed as a basic approximation element and error
back propagation algorithm [46, 47] by fixing hidden nodes as per
the required accuracy. The trial solution of the model is generated
by training the algorithm. The approximate solution by AN N has
numerous benefits compared with traditional numerical methods.
The neural network trial solution is written as sum of two terms;
the first one satisfies initial/boundary conditions and the second
part involves neural network with adjustable parameters.
3.
DESCRIPTION OF THE METHOD
The generic form of sixth order differential equations may be represented as

d6 y(x)
dy d2 y d3 y d4 y d5 y
=
f
(x,
y,
,
,
,
,
) x [a, b]
dx6
dx dx2 dx3 dx4 dx5
subject to boundary condition
y(a) = A;
00
y iv (a) = E
00
y iv (b) = F
y (a) = C;
y(b) = B;
y (b) = D;
(1)
(2)
Assuming that the trial solution is of the form

yT (x, p) = J(x) + M (x)N (x, p)
(3)
There are any forms for functions J(x) and M (x) to construct
yT (x, p). Sigmoid function tanh(bi + ci x) is accepted to be an
activation function for hidden unit. The trial function can not be
satisfied in the boundary conditions unless there is
00
J iv (a) = E
00
J iv (b) = F
J(a) = A;
J (a) = C;
J(b) = B;
J (b) = D;
(4)
The function M (x) is taken as (x a)3 (x b)3 . This motivates

the introduction of J(x) = wo + w1 x + w2 x2 + ... + w5 x5 which
is polynomial of five degree with coefficient wo , w1 , ...w5
00
M (a).N (a, p) = 0
(5)
M (b).N (b, p) = 0
(6)
00
00
M (a).N (a, p) + 2M (a).N (a, p) + M (a).N (a, p) = 0 (7)

00
M (b).N (b, p) + 2M (b).N (b, p) + M (b).N (b, p) = 0 (8)

000
00
00
M iv (a).N (a, p) + 4M (a).N (a, p) + 6M (a).N (a, p)+

0
000
4M (a).N (a, p) + M (a).N iv (a, p) = 0

(9)
000
00
00
M iv (b).N (b, p) + 4M (b).N (b, p) + 6M (b).N (b, p)+

0
000
4M (b).N (b, p) + M (b).N iv (b, p) = 0

(10)
The following conditions must be satisfied with Eq. 3
wo + w1 a + w2 a2 + w3 a3 + w4 a4 + w5 a5 = A
(11)

Table 1. Values of Neural parameters
(Example 1), ai , bi and ci with five nodes
i
1
2
3
4
5
ai
0.016505561
0.087265869
0.024116433
0.135271789
-0.19112553
bi
-0.2434469
-0.0058665
-0.0300997
-0.0075984
-0.1365936
Table 2. Error Estimation of Example 1

x
ci
0.863716
1.069340
1.0776682
0.7272999
1.4643935
Numerical
Solution
1.0000000000
0.9908587500
0.9699309700
0.9350350300
0.8835115500
0.8121660400
0.7172044500
0.5941605000
0.4378140300
0.2420990000
0.0000000000
Error
0.0000000000
0.0037950800
0.0071912400
0.0098661300
0.0115832700
0.0121945900
0.0116430700
0.0099653100
0.0072941500
0.0038613100
0.0000000000
wo + w1 b + w2 b2 + w3 b3 + w4 b4 + w5 b5 = B
(12)
2w2 + 6w3 a + 12w4 a2 + 20w5 a3 = C
(13)
2w2 + 6w3 b + 12w4 b2 + 20w5 b3 = D
(14)
Fig. 1. Graph of exact solution in comparison with the computed solution

of Example 1
24w4 + 120w5 a = E
(15)

24w4 + 120w5 b = F
(16)
After solving this system of six equations with six unknowns, the
general form of the polynomial J(x) is achieved. For minimization,
the following equation is used
E(p) =
M
X
d6 yT (xi , p)
dyT (xi , p)
,
f (xi , yT (xi , p),
{
6
dx
dx
i=1
(17)
d2 yT (xi , p)
d5 yT (xi , p) 2
.....
)}
2
dx
dx5
3.1
0.0000000000
0.1000000000
0.2000000000
0.3000000000
0.4000000000
0.5000000000
0.6000000000
0.7000000000
0.8000000000
0.9000000000
1.0000000000
Analytical
Solution
1.0000000000
0.9946538300
0.9771222100
0.9449011700
0.8950948200
0.8243606400
0.7288475200
0.6041258100
0.4451081900
0.2459603100
0.0000000000
3.1.1 Example 1:. Let us shed some light on the sixth-order differential equation:
0<x<1
(18)
y(1) = 0;
00
00
y (1) = 2e;
ci
-0.2868379
0.5234077
0.3504931
-0.2325288
0.7076312
3.1.2 Example 2. Consider the following sixth order differential

equation:
y(0) = 1;
0<x<1
y iv (0) = 3
y iv (1) = 4e
(22)
y(1) = e;
y iv (1) = e
y (1) = e;
(23)
The ANN trial solution is written as

yT (x, p) = wo + w1 a + w2 a2 + w3 a3 + w4 a4 + w5 a5
1
,
24
(24)
5e8
,
36
(20)
(21)
3.1.3 Example 3. The boundary value problem may be consider

as follows:
(19)
The trial solution has been obtained as

y iv (0) = 1
00
where w5 =
w4 =
w3 =
w2 = 12, w1 =
307e472
and
w
=
1.
The
values
of
Neural
parameters
ai , bi and
o
360
ci , where i represents numbers of nodes, are shown in Table 3. The
network is trained for 10 equidistant points with five hidden nodes
according to the algorithm. In Table 4, the analytical solutions are
compared with neural solutions. Also estimated errors are cited in
third column of Table 4. Plot of comparison between (analytical
results) and ( neural results) of Example 2, is shown in Figure 2.
The exact solution of above problem is

y(x) = (1 x)ex
00
y (0) = 1;
e1
,
120

y (0) = 1;
bi
-0.6258868
-0.6391466
-0.5823448
-0.6028480
-0.4779116
In this section, the solution of three example problems with the help
of neural network method is presented
y(0) = 1;
ai
-0.0734690
0.0151022
-0.0296600
0.0867887
0.0018519
y vi (x) = 6ex + y 2 (x)
Numerical Implementation
y vi (x) = 6ex + y(x)
i
1
2
3
4
5
, w4 = 18 , w3 = 32e
, w2 = 12,
where w5 = 34e
120
9
23e66
w1 = 15 and wo = 1. The values of Neural parameters ai ,
bi and ci , where i represents numbers of nodes, are shown in Table
1. The network has been trained for 10 equidistant points in [0, 1]
and compared results between analytical and neural with five nodes
fixed in the hidden layer. Comparison between analytical and neural results is outlined in Table 2. Error between two values is also
computed here. Analytical and neural results of Example 1, which
are obtained are depicted in Figure 1.
y vi (x) = ex y 2 (x)
0<x<1
(25)

y(0) = 1;
y(1) = e1 ;
y (0) = 1;
00
y (0) = 1
y (1) = e1 ;
00
y (1) = e1
(26)
The exact solution of the problem is

y(x) = ex
(27)

x
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
Analytical
Solution
1.00000000
0.90483742
0.81873075
0.74081822
0.67032005
0.60653066
0.54881164
0.49658530
0.44932896
0.40656966
0.36787944
Numerical
Solution
1.00000000
0.90483742
0.81873075
0.74081822
0.67032005
0.60653066
0.54881164
0.49658530
0.44932896
0.40656966
0.36787944
Error
0.00000000
0.00000000
0.00000000
0.00000000
0.00000000
0.00000000
0.00000000
0.00000000
0.00000000
0.00000000
0.00000000
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1

of Example 2
i
1
2
3
4
5
ai
1.01544450
1.23698750
0.09458130
0.00376860
0.09513450
bi
0.07528380
0.07341860
0.07306020
0.14662330
0.07199000
Numerical
Solution
1.00000000
1.10585595
1.22270049
1.35163868
1.49391357
1.65091952
1.82421682
2.01554776
2.22685442
2.46029828
2.71828183
Error
0.00000000
-0.00068503
-0.00129774
-0.00177987
-0.00208887
-0.00219825
-0.00209802
-0.00179505
-0.00131349
-0.00069517
0.00000000

of Example 3
5.
Analytical
Solution
1.00000000
1.10517092
1.22140276
1.34985881
1.49182470
1.64872127
1.82211880
2.01375271
2.22554093
2.45960311
2.71828183
CONFLICT OF INTERESTS
The authors declare that there is no conflict of interests regarding

the publication of this paper.
ci
0.12410432
0.13873990
0.26252581
0.50823108
0.17995931
6.
ACKNOWLEDGMENT
We thank the reviewers for their thorough efforts in editing our paper and highly appreciate the comments and constructive criticism
that significantly contributed in improving the quality of the publication. The authors also thank Ms. Wishaal Khalid for proofreading
our research paper.
The ANN trial solution is written as

(28)
1
1
1
, w3 = 29e 2 11 , w2 = 12,
w5 = 19e 2 7 , w4 = 1746e
2
1 and wo = 1. Table 5 gives the values of Neural param-
where
w1 =
eters ai , bi and ci , at different nodes. Ten equidistant points are
taken into consideration in the given domain at five hidden nodes.
Comparison of analytical and neural results of Example 3 has been
shown in Table 6.
It is well known that other numerical methods are usually iterative
in nature, where the step size is fixed before the commencement of
the computation. After the attainment of the solution if we decide
to know what it is between steps, there is no option but to repeat
the procedure from initial stages. Neural Network method may be
a sort of reliefs we bring to this problem, where this repetition of
iterations may be avoided. One can see from the tables and graphs
that the trial solution approach improves the accuracy of the results.
Lastly, it would not go to amiss to mention that the implemented algorithm is simple, computationally efficient, and straight forward.
4.
CONCLUDING REMARKS
In this paper, a stable and efficient feed forward neural network

method for sixth-order differential equations with boundary conditions. Numerical results are presented that exhibit the high accuracy
of the proposed algorithms. The paper may be concluded on the
note that this method is extremely powerful, it may be employed
in future calculations by researchs in various filed of engineering.
The mian application of the technique in question lies in finding
solutions for a higher order of differential eqution.
7.
REFERENCES
[1] Schalkoff, R.J. (1997) Artificial Neural Networks.
McGraw-Hill, New York.
[2] Picton, P. (2000) Neural Networks. 2nd ed. Palgrave, Great
Britain.
[3] Rosenblatt, F. (1957) The perceptron - a perceiving and
recognizing automation. Comell Aeronautical Laboratory,
Report 85460.
[4] Boutayeb, A. & Twizell, E. (1992) Numerical methods for
the solution of special sixth-order boundary value problems.International Journal of Computer Mathematics,45.
pp 207-233
[5] Twizell, E. & Boutayeb, E. (1990) Numerical methods for
the solution of special and general sixth-order boundary
value problems, with applications to Benard layer eigenvalue problems. Proceedings of the Royal Socity London
Series A. 431(1883). pp 433-450
[6] Baldwin, P. (1987) Asymptotic estimates of the eigenvalues
of a sixth-order boundary-value problem obtained by using
global phase-integral methods. Philosophical Transactions
of the Royal Society of London Series A, 322(1566). pp
281-305
[7] Siddiqi, S.S. & Twizell, E.H. (1996) Spline solutions of
linear sixth-order boundary-value problems. International
Journal of Computer Mathematics, 60(3-4). pp 295-304
[8] El-Gamel, M., Cannon, J.R. & Zayed, A.I. (2004) SincGalerkin method for solving linear sixth-order boundaryvalue problems. Mathematics of Computation, 73(247). pp
1325-1343

[9] Agarwal, R.P. (1986) Boundary Value Problems for Higher
Order Differential Equations. World Scientific Publishing,
Singapore
[10] Shen, J. (1994) Efficient spectral-Galerkin method. I. Direct solvers of second- and fourth-order equations using
Legendre polynomials. SIAM Journal on Scientific Computing, 15(6). pp 1489-1505
[11] Doha, E.H. & Bhrawy, A.H. (2006) Efficient spectralGalerkin algorithms for direct solution for second-order
differential equations using Jacobi polynomials. Numerical
Algorithms, 42(2). pp 137-164
[12] Doha, E.H. & Bhrawy, A.H. (2008) Efficient spectralGalerkin algorithms for direct solution of fourth-order differential equations using Jacobi polynomials. Applied Numerical Mathematics, 58(8). pp 1224-1244
[13] Doha, E.H. & Bhrawy, A.H. (2009) A Jacobi spectral
Galerkin method for the integrated forms of fourth-order
elliptic differential equations. Numerical Methods for Partial Differential Equations, 25(3).pp 712-739
[14] Doha, E.H., Bhrawy, A.H. & Abd-Elhameed, W.M. (2009)
Jacobi spectral Galerkin method for elliptic Neumann
problems. Numerical Algorithms, 50(1). pp 67-91
[15] Doha, E.H., Bhrawy, A.H. & Hafez, R.M. (2011) A JacobiJacobi dual-Petrov-Galerkin method for third- and fifthorder differential equations. Mathematical and Computer
Modelling, 53(9-10). pp 1820-1832
[16] Doha, E.H., Bhrawy, A.H. & Hafez, R.M. (2011) A Jacobi
dual-Petrov-Galerkin method for solving some odd-order
ordinary differential equations. Abstract and Applied Analysis, Artical ID 947230, 21 pages
[17] Bhrawy, A.H. & Abd-Elhameed, W.M. (2011) New algorithm for the numerical solutions of nonlinear thirdorder differential equations using Jacobi-Gauss collocation method. Mathematical Problems in Engineering, Article ID 837218, 14 pages
[18] Doha, E.H., Bhrawy, A.H. & Saker, M.A. (2011) Integrals
of Bernstein polynomials: an application for the solution of
high even-order differential equations. Applied Mathematics Letters, 24(4). pp 559-565
[19] Doha, E.H., Bhrawy, A.H. & Saker, M.A. (2011) On
the derivatives of Bernstein polynomials: an application
for the solution of high even-order differential equations.
Boundary Value Problem, Artical ID 829543, 16 pages
[20] Loghmani, G.B. & Ahmadinia, M. (2007) Numerical solution of sixth order boundary value problems with sixth
degree B-spline functions. Applied Mathematics and Computation, 186(2). pp 992-999
[21] Siddiqi, S.S. & Akram, G. (2008) Septic spline solutions of
sixth-order boundary value problems. Journal of Computational and Applied Mathematics, 215(1). pp 288-301
[22] Wazwaz, A. (2001) The numerical solution of sixth-order
boundary value problems by the modified decomposition
method. Applied Mathematics and Computation, 118(2-3).
pp 311-325
[23] Bhrawy, A.H. (2009) Legendre-Galerkin method for sixthorder boundary value problems. Journal of the Egyptian
Mathematical Society, 17(2). pp 173-188
[24] Lee. H. & Kang, I.S. (1990) Neural algorithm for solving
differential equations. Journal of Computational Physics,
91(1). pp 110-131
[25] Meade, A.J. & Femandez, A.A. (1994) The numerical solution of linear ordinary differential equations by feedforward neural networks. Mathematical and Computer Modelling, 19(12). pp 1-25
[26] Meade, A.J. & Femandez, A.A. (1994) Solution of nonlinear ordinary differential equations by feedforward neural
networks. Mathematical and Computer Modelling, 20(9).
pp 19-44
[27] Lagaris, I.E., Likas, A. & Fotiadis, D.I. (1998) Artificial
neural networks for solving ordinary and partial differential equations. IEEE Transactions on Neural Networks,
9(5) pp 987-1000
[28] Lagaris, I.E., Likas, A. & Papageorgiou, D.G. (2000)
Neural-network methods for boundary value problems with
irregular boundaries. IEEE Transactions on Neural Networks, 11(5). pp 1041-1049
[29] Parisi, D.R., Mariani, M.C. & Laborde, M.A. (2003) Solving differential equations with unsupervised neural networks. Chemical Engineering and Processing, 42(8-9). pp
715-721
[30] Malek, A. & Shekari, B. (2006) Numerical solution for
high order differential equations using a hybrid neural
network-Optimization method. Applied Mathematics and
Computation, 183(1). pp 260-271
[31] Choi, B. & Lee, J.H. (2009) Comparison of generalization
ability on solving differential equations using backpropagation and reformulated radial basis function networks.
Neurocomputing, 73(1-3). pp 115-118
[32] Yazdi, H.S., Pakdaman, M. & Modaghegh, H. (2011) Unsupervised kernel least mean square algorithm for solving
ordinary differential equations. Neurocomputing, 74(1213). pp 2062-2071
[33] Selvaraju, N. & Samant, A. (2010) Solution of matrix Riccati differential equation for nonlinear singular system using neural networks. International Journal of Computer
Applications, 29. pp 48-54
[34] He, S., Reif, K. & Unbehauen, R. (2000) Multilayer neural
networks for solving a class of partial differential equations. Neural Networks, 13(3). pp 385-396
[35] Kumar, M. & Yadav, N. (2011) Multilayer perceptrons and
radial basis function neural network methods for the solution of differential equations: a survey. Computers and
Mathematics with Applications, 62(10). pp 3796-3811
[36] Tsoulos, I.G., Gavrilis, D. & Glavas, E. (2009) Solving differential equations with constructed neural networks. Neurocomputing, 72(10-12). pp 2385-2391
[37] Jianyu, L., Siwei, L, Yingjian, Q. & Yaping, H. (2003)
Numerical solution of elliptic partial differential equation
using radial basis function neural networks. Neural Networks, 16(5-6). pp 729-734
[38] Shirvany, Y., Hayati, M. & Moradian, R. (2009) Multilayer
perceptron neural networks with novel unsupervised training method for numerical solution of the partial differential
equations. Applied Soft Computing Journal, 9(1). ppp 2029
[39] Mai-Duy, N. & Tran-Cong, T. (2001) Numerical solution of
differential equations using multiquadric radial basis function networks. Neural Networks, 14(2). pp 185-199

[40] Leephakpreeda, T. (2002) Novel determination of
differential-equation solutions: universal approximation method. Journal of Computational and Applied
Mathematics, 146(2). pp 443-457
[41] Franke, C. & Schaback, R. (1998) Solving partial differential equations by collocation using radial basis functions.
Applied Mathematics and Computation, 93(1). pp 73-82
[42] Smaoui, N. & Al-Enezi, S. (2004) Modelling the dynamics of nonlinear partial differential equations using neural
networks. Journal of Computational and Applied Mathematics, 170(1). pp 27-58
[43] Tsoulos, I.G. & Lagaris, I.E. (2006) Solving differential
equations with genetic programming. Genetic Programming and Evolvable Machines, 7(1). pp 33-54
[44] McFall, K.S. & Mahan, J.R. (2009) Artificial neural network method for solution of boundary value problems with
exact satisfaction of arbitrary boundary conditions. IEEE
Transactions on Neural Networks, 20(8). pp 1221-1233
[45] Hoda, S.A. & Nagla, H.A. (2011) Neural network methods
for mixed boundary value problems. International Journal
of Nonlinear Science, 11. pp 312-316
[46] Zurada, J.M. (1994) Introduction to Artificial Neural Network.West Publishing.
[47] Haykin, S. (1999) Neural Networks a Comprehensive
Foundation. Prentice Hall, New York.

Numerical Solution of Sixth-Order Differential Equations Arising in Astrophysics by Neural Network

Uploaded by

Copyright:

Available Formats

Numerical Solution of Sixth-Order Differential Equations Arising in Astrophysics by Neural Network

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Numerical Solution of Sixth-Order Differential Equations Arising in Astrophysics by Neural Network

Uploaded by

Copyright:

Available Formats

International Journal of Computer Applications (0975 8887)

Volume 107 - No. 6, December 2014