Anda di halaman 1dari 19

Mediterr. J. Math.

13 (2016), 26132631
DOI 10.1007/s00009-015-0642-z
1660-5446/16/052613-19
published online October 17, 2015
c Springer Basel 2015

Second Kind Chebyshev Wavelet Galerkin


Method for Stochastic It
o-Volterra Integral
Equations
Fakhrodin Mohammadi

Abstract. In this paper, an ecient wavelet Galerkin method based on


the stochastic operational matrix of second kind Chebyshev wavelet is
proposed for solving stochastic It
o-Volterra integral equations. Conver-
gence and error analysis of the presented wavelets method are inves-
tigated. The numerical results are compared with exact solution and
those of other existing methods.
Mathematics Subject Classification. 65T60, 45L05, 60H20.
Keywords. Second kind Chebyshev wavelets, It o integral, Stochastic
operational matrix, Stochastic It
o-Volterra integral equations.

1. Introduction
Recently, the topic of stochastic dierential and integral equations has been
an area of active research in mathematics, uid mechanics, geophysics, biol-
ogy, chemistry, epidemiology, microelectronics, theoretical physics, econom-
ics, and nance. The behavior of dynamical systems in these areas is often
dependent on a noise source and a Gaussian white noise, governed by cer-
tain probability laws, so that modeling such phenomena naturally requires
the use of various stochastic dierential equations or, in more complicated
cases, stochastic integral equations and stochastic integro-dierential equa-
tions. Since analytic solutions of stochastic integral and dierential equations
are not available in many cases, numerical approximation becomes a prac-
tical way to face this diculty. Many studies have been carried out on the
problem of approximating the solution of stochastic integral and dierential
equations [114].
In the recent years, many orthogonal basis functions, such as block
pulse functions, Walsh functions, Fourier series, orthogonal polynomials and
wavelets, were used to estimate solution of functional equations. As a pow-
erful tool, wavelets have found their way into many dierent elds of science
2614 F. Mohammadi MJOM

and engineering. Wavelets permit the accurate representation of a variety of


functions and operators. Moreover, wavelets establish a connection with fast
numerical algorithms [1517]. Several wavelets methods have been used for
approximating solution of the integral equations and dierential equations
[1826].
Recently, numerical solution of stochastic functional equations becomes
an interesting subject in numerical analysis [414]. In this paper, we consider
the following stochastic Volterra integral equation
t t
X(t) = f (t) + a (s, X(s)) ds + (s, X(s)) dB(s), t [0, T ), (1)
0 0

where X(t), f (t), a (t, X(t)) and (t, X(t)), for t [0, T ), are stochas-
tic processes dened on the same probability space (, F, P ), and X(t) is
t
unknown. Also B(t) is a Brownian motion process and 0 (s, X(s)) dB(s)
is the It
o integral. Many researchers study the numerical solution of this kind
of stochastic integral equation. The block pulse functions and their stochas-
tic operational matrix are used for solving stochastic It o-Volterra integral
equations in [4]. Khodabin et al. [6] have proposed a computational method
based on Triangular functions and their operational matrix for solving sto-
chastic Ito-Volterra integral equations. In [7] and [8], the generalized hat basis
functions and their operational matrix are used for numerical solution of sto-
chastic Ito-Volterra integral equations. Mohammadi [9] has been presented
a computational method based on the Chebyshev wavelets for solving sto-
chastic Ito-Volterra integral equations. Asgari et al. [11] have used Bernstein
polynomials and their operational matrices for solving nonlinear stochastic
integral equation. To compute the approximate solution of stochastic It o-
Volterra integral Eq. (1) the second kind Chebyshev wavelets are introduced,
then an stochastic operational matrix for these orthogonal wavelets is derived.
Using this stochastic operational matrix a computational method is proposed
for solving stochastic Ito-Volterra integral equations.
The rest of this paper is organized as follows: in Sect. 2 some basic def-
initions and preliminaries about stochastic process and It o integral are pre-
sented. The second kind Chebyshev wavelets and their properties are intro-
duced in Sect. 3. In Sect. 4 stochastic operational matrix of the second kind
Chebyshev wavelets are derived. Section 5 is devoted to applications of this
stochastic operational matrix in solving stochastic It o-Volterra integral equa-
tions. Convergence analysis of the proposed method is considered in Sect. 6.
Numerical examples are included in Sect. 7. Finally, a conclusion is given in
Sect. 8.

2. Preliminaries
In this section, we review some basic denition of the stochastic calculus and
the block pulse functions (BPFs).
Vol. 13 (2016) Second Kind Chebyshev Wavelet 2615

2.1. Stochastic Calculus


Definition 1 (Brownian motion process). A real-valued stochastic process
B(t), t [0, T ] is called Brownian motion, if it satises the following proper-
ties
(i) The process has independent increments for 0 t0 t1 tn T ,
(ii) For all t 0, B(t + h) B(t) has Normal distribution with mean 0 and
variance h,
(iii) The function t B(t) is continuous functions of t.
Definition 2. Let {Nt }t0 be an increasing family of -algebras of subsets of
. A process g(t, ) : [0, ) Rn is called Nt -adapted if for each t 0
the function g(t, ) is Nt -measurable.
Definition 3. Let V = V(S, T ) be the class of functions f (t, ) : [0, )
R such that
(i) The function (t, ) f (t, ) is B F-measurable, where B denotes the
Borel algebra on [0, ) and F is the -algebra on .
(ii) f is adapted to Ft , where Ft is the -algebra generated by the random
variables
 B(s), s  t.
T 2
(iii) E S f (t, )dt < .

o integral). Let f V(S, T ), then the It


Definition 4 (The It o integral of f is
dened by
T T
f (t, )dBt () = lim n (t, )dBt (), (lim in L2 (P ))
n
S S

where, {n } is a sequence of elementary functions such that


T

2
E (f (t, ) n (t, )) dt 0, asn .
s

o isometry). Let f V(S, T ), then


Definition 5 (The It
2 T
T 

E f (t, )dB() = E f 2 (t, )dt .
S S

For more details about stochastic calculus please see [2].

2.2. Block Pulse Functions


BPFs have been studied by many authors and applied for solving dierent
problems. In this section, we recall denitions and some properties of the
block pulse functions [4,30].
The m-set of BPFs are dened as

1 (i 1)h t < ih,
bi (t) =
0 otherwise,
2616 F. Mohammadi MJOM

T
in which t [0, T ), i = 1, 2, . . . , m and h = m. The BPFs are disjointed with
each other in the interval [0, T ) and
bi (t)bj (t) = ij bi (t), i, j = 1, 2, . . . , m,
where ij is the Kronecker delta. The set of BPFs dened in the interval
[0, T ) are orthogonal with each other, that is
T
bi (t)bj (t)dt = hij , i, j = 1, 2, . . . , m.
0

If m the set of BPFs is a complete basis for L2 [0, T ), so an arbitrary


real bounded function f (t), which is square integrable in the interval [0, T ),
can be expanded into a block pulse series as
m

f (t)  fi bi (t), (2)
i=1

where
T
1
fi = bi (t)f (t)dt, i = 1, 2, . . . , m.
h
0

Rewriting Eq. (2) in the vector form we have


m

f (t)  fi bi (t) = F T (t) = T (t)F, (3)
i=1

in which
T T
F = [f1 , f2 , . . . , fm ] , (t) = [b1 (t), b2 (t), . . . , bm (t)] .
Moreover, any two-dimensional function k(s, t) L2 ([0, T1 ] [0, T2 ]) can be
expanded with respect to BPFs such as
k(s, t) = T (t)(t),
where (t) is the m-dimensional BPFs vector, respectively, and is the
m m BPFs coecient matrix with (i, j)th element
T1 T2
1
ij = k(s, t)bi (t)bj (s)dtds, i, j = 1, . . . , m
h1 h2
0 0
T1 T2
and h1 = m and h2 = m. Let (t) be the BPFs vector, then we have
T (t)(t) = 1,
and

b1 (t) 0 ... 0
.. ..
0 b2 (t) . .
(t)T (t) =
.

.
.. .. ..
. . 0
0 ... 0 bm (t) mm
Vol. 13 (2016) Second Kind Chebyshev Wavelet 2617

For an m-vector F we have


(t)T (t)F = F (t), (4)
where F is an m m matrix, and F = diag(F ).

3. Second Kind Chebyshev Wavelets


Wavelets constitute a family of functions constructed from dilation and trans-
lation of a single function called the mother wavelet. When the dilation
parameter a and the translation parameter b vary continuously, we have the
following family of continuous wavelets
 
12 tb
a,b (t) = a , a, b R, a = 0.
a
The second kind Chebyshev wavelets nm (x) are dened on the interval [0, 1)
by [2729]
 k +1  
22

Um 2k+1 t 2n 1 , 2nk t n+1
2k
nm (t) =
0, otherwise
where Um (t) is the second kind Chebyshev polynomials of degree m, given
by [31]
sin ((m + 1))
Um (t) = , t = cos().
sin()
The second kind Chebyshev wavelets {nm (t)|n = 0, 1, . . . , 2k 1, m =
0, 1, 2, . . . , M 1} forms an orthonormal basis for L2wnk [0, 1] with respect to
 
the weight function wnk (t) = w 2k+1 t 2n 1 , in which w(t) = 1 t2 .
Using the orthonormality of the second kind Chebyshev wavelets, any
function f (t) over [0, 1); square integrable with respect to the measure w(t)dt;
with w(t) = wnk (t); for 2nk t n+1 ; and wnk (t) = w(2k+1 t 2n + 1); being
2k
w(t) = 1 t2 can be expanded in terms of the second kind Chebyshev
wavelets as


f (t)  cmn mn (t), (5)
n=0 m=0

where cmn = (f (t), mn (t))wnk and (., .)wnk denotes the inner product on
L2wnk [0, 1]. If the innite series in (5) is truncated, then it can be written as
k
2 1 M
 1
f (t)  cmn mn (x) = C T (t), (6)
n=0 m=0

where C and (t) are m = 2k M column vectors given by


 T
C = c00 , . . . , c0(M 1) , c10 . . . , c1(M 1) , . . . , c(2k 1)(M 1) ,
 T
(x) = 00 (t), . . . , 0(M 1) (t), 10 (t) . . . , 1(M 1) (t), . . . , (2k 1)(M 1) (t) .
2618 F. Mohammadi MJOM

By changing indices in the vectors (t) and C the series (6) can be rewritten
as
m


f (t)  ci i (t) = C T (t), (7)
i=1

where
T T
C = [c1 , c2 , . . . , cm
] , (x) = [1 (x), 2 (x), . . . , m
(x)] , (8)
and
ci = cnm , i (t) = nm (t), i = (n 1)M + m + 1.
Similarly, any two-dimensional function k(s, t) L2w  w ([0, 1] [0, 1]) can be
expanded into second kind Chebyshev wavelets basis as
m

 m

k(s, t) kij i (s)j (t) = T (s)K(t), (9)
i=1 j=1
 
m
where K = [kij ] is an m matrix and kij = i (s), (k(s, t), j (t))wnk .
wnk

3.1. Second Kind Chebyshev Wavelets and BPFs


In this section, we will review the relation between the second kind Chebyshev
wavelets and BPFs. It is worth mention that here we set T = 1 in denition
of BPFs.
Theorem 3.1. Let (t) and (t) be the m-dimensional
second kind Chebyshev
wavelets and BPFs vector, respectively, the vector (t) can be expanded by
BPFs vector (t) as
(t)  Q(t), (10)
m
where Q is an m block matrix and
 
2j 1
Qij = i , i, j = 1, 2, . . . , m.

2m
Proof. Let i (t), i = 1, 2, . . . , m
be the ith element of second kind Chebyshev
wavelets vector. Expanding i (t) into an m-term vector of BPFs, we have
m


i (t)  Qik bk (t), i = 1, 2, . . . , m,
(11)
k=1

2j1
taking the collocation points j = 2m
and evaluating relation (11) we get
m


i (j )  Qik bk (j ) = Qij , i, j = 1, 2, . . . , m,
(12)
k=1

and this prove the desired result. 


Vol. 13 (2016) Second Kind Chebyshev Wavelet 2619

3.2. Convergence Analysis


Here, we investigate the convergence and error analysis of the second kind
Chebyshev wavelets basis.
Theorem 3.2. Supposef (t) L2w [0, 1] with bounded second derivative, say

|f (t)| L, and let n=0 m=0 cmn mn (t) be its innite second kind Cheby-
shev wavelets expansion, then

L
|cmn | 5 , (13)
23 (n + 1) 2 (m2 + 2m 3)
this means the second kind Chebyshev wavelets series converges uniformly to
f (x) and


f (t) = cnm nm (t),
n=1 m=0

Proof. From the denition of coecient cmn we have


1
cmn = f (t)mn (t)wnk (t)dt
0

n+1
k 2k
2 2 +1
= f (t)Um (2k+1 t 2n 1)wnk (t)dt,

n
2k

by substituting 2k+1 t 2n 1 = cos() we get


k   
2 2 cos() + 2n + 1
cmn = f sin (m + 1) sin()d
2k+1
0

using integration by part two times we obtain


  
1  cos() + 2n + 1
cmn = f hm ()d
2
5k
2 +3 2k+1
0

where
 
sin() sin ((m 1)) sin ((m + 1))
hm () =
m m1 m+1
 
sin() sin ((m + 1)) sin ((m + 3))
,
m+2 m+1 m+3
so, we have
5k  5k
L2( 2 3) L2( 2 3)
|cmn | |hm ()| d
(m2 + 2m 3)
0
2620 F. Mohammadi MJOM

since n 2k 1, we obtain

L
|cmn | 5k
2 2 +3 (m2 + 2m 3)

L
5 .
23 (n + 1) 2 (m2 + 2m 3)

Theorem 3.3. Let f (t) be a continuous function dened on [0, 1), with second
derivatives f  (t) bounded by L, then we have the following accuracy estima-
tion

L2   1
M,k 6 5 2
2 n=0 2
(n + 1) (m + 2m 3)
m=M

M 1
 12
L2   1
+ 6 5 2 ,
2
k m=0 (n + 1) (m + 2m 3)
2
n=2
where
2 12
1 k
2 1 M
 1
f (t)
M,k = cnm nm (t) w(t)dt .
0 n=0 m=0

Proof. We have
2
1 k
2 1 M
 1
2
M,k = f (t) cnm nm (t) w(t)dt
0 n=0 m=0
2
1 
 k
2 1 M
 1
= cnm nm (t) cnm nm (t) w(t)dt
0 n=0 m=0 n=0 m=0


 1 M
  1 1
= c2nm 2
nm (t)w(t)dt + c2nm 2
nm (t)w(t)dt
n=0 m=M 0 n=2k m=0 0


 M
  1
= c2nm + c2nm ,
n=0 m=M n=2k m=0
now by considering the relation (13) we achieve the desired result. 

4. Stochastic Operational Matrix


In this section, we derive an stochastic operational matrix for the second kind
Chebyshev wavelets. For this purpose, we rst remind some useful results for
BPFs[4,5].
Lemma 4.1. Let (t) be the m-dimensional
BPFs vector dened in (3), then
integration of this vector can be derived as
Vol. 13 (2016) Second Kind Chebyshev Wavelet 2621

t
(s)ds  P (t), (14)
0
where P is called the operational matrix of integration for BPFs and is given
by

1 2 2 ... 2
0 1 2 ... 2

h .. ..
P = 0 0 1 . . .
2. .. .. ..


.. . . . 2
0 0 0 ... 1 m m

Proof. Please see [4]. 
Lemma 4.2. Let (t) be the m-dimensional
BPFs vector dened in (3), the
It
o integral of this vector can be derived as
t
(s)dB(s)  Ps (t), (15)
0
where Ps is called the stochastic operational matrix of BPFs and is given by

1 1 1 ... 1
0 2 1 ... 1

. ..

Ps = 0 0 3 . . . .
. . . .
.. .. .. ..
m1


0 0 ... 0 m
 
(2i1)h
where i = B 2 B ((i 1)h) and i = B (ih) B ((i 1)h).
Proof. Please see [4]. 
Now we are ready to derive a new operational matrix of stochastic inte-
gration for the second kind Chebyshev wavelets basis. For this end we use
BPFs and the matrix Q introduced in (10).
Theorem 4.1. Suppose (t) be the m-dimensional
second kind Chebyshev
wavelets vector dened in (7), the integral of this vector can be derived as
t
(s)ds  QP Q1 (t) = (t), (16)
0
where Q is introduced in (10) and P is the operational matrix of integration
for BPFs derived in (14).
Proof. Let (t) be the second kind Chebyshev wavelets vector, using Theo-
rem 3.1 and Lemma 4.1 we have
t t t
(s)ds  Q(s)ds =Q (s)ds = QP (t),
0 0 0
2622 F. Mohammadi MJOM

now Theorem 3.1 gives


t
(s)ds QP (t) = QP Q1 (t) = (t),
0

and this completes the proof. 

Theorem 4.2. Suppose (t) be the m-dimensional


second kind Chebyshev
wavelets vector dened in (7), the It
o integral of this vector can be derived as
t
(s)dB(s)  QPs Q1 (t) = s (t), (17)
0

where s is called stochastic operational matrix for second kind Chebyshev


wavelets, Q is introduced in (10) and Ps is the stochastic operational matrix
of integration for BPFs derived in (15).
Proof. Let (t) be the second kind Chebyshev wavelets vector, using Theo-
rem 3.1 and Lemma 4.2 we have
t t
(s)dB(s)  Q(s)dB(s)
0 0
t
=Q (s)dB(s) = QPs (t),
0

now Theorem 3.1 results


t
(s)dB(s) =QPs (t) = QPs Q1 (t) = s (t),
0

and this completes the proof. 

5. Method of Solution
In this section, we solve stochastic Ito-Volterra integral equations using the
stochastic operational matrix of the second kind Chebyshev wavelets. Con-
sider the following stochastic It
o-Volterra integral equation as
t t
X(t) = f (t) + a (s, X(s)) ds + (s, X(s)) dB(s), t [0, T ),
0 0

where X(t), f (t), a (t, X(t)) and (t, X(t)), for t [0, T ), are stochas-
tic processes dened on the same probability space (, F, P ), and X(t) is
t
unknown. Also B(t) is a Brownian motion process and 0 (s, X(s)) dB(s)
is the Ito integral. For solving this problem using the stochastic opera-
tional matrix of second kind Chebyshev wavelets, we approximate X(t), f (t),
Vol. 13 (2016) Second Kind Chebyshev Wavelet 2623

a (t, X(t)) and (t, X(t)) in terms of m-dimensional


second kind Chebyshev
wavelets as follows
f (t) = F T (t) = T (t)F, (18)

X(t) = X T (t) = T (t)X, (19)

a (t, X(t)) = K1T (t) = (t)T K1 , (20)

(t, X(t)) = K2T (t) = (t)T K2 , (21)


where X, F, K1 and K2 are second kind Chebyshev wavelet coecients
vector dened in Eq. (8). Moreover, K1 and K2 are functions of the unknown
vector X. Substituting above approximations in Eq. (1), we have
t t
T T
X (t) = F (t) + K1T (s)ds + K2T (s)dB(s), (22)
0 0
applying the operational matrices and s for second kind Chebyshev
wavelets derived in Eq. (22) we get
X T (t) = F T (t) + K1T (t) + K2T s (t), (23)
This equation is held for all t [0, 1), so we can write
X T K1T K2T F T = 0. (24)
Since K1 and K2 are functions of X, Eq. (24) is a nonlinear system of equa-
tions for the unknown vector X. This nonlinear system can be solved using the
well-known NewtonRaphson method. After solving this system and deter-
mining vector X, we can approximate solution of the stochastic Ito-Volterra
integral Eq. (1) by substituting obtained vector X in Eq. (19).

6. Convergence Analysis
In this section, we consider the convergence properties of the second kind
Chebyshev wavelet method for solving the stochastic It o-Voltera integral Eq.
(1). Now let . 2 be the L2w norm on [0, 1] and em (t) = X(t) Xm (t) be an
error function of approximate solution Xm (t) to the exact solution X(t).
Theorem 6.1. Let X(t) be exact solution and Xm (t) be the second kind Cheby-
shev wavelet approximate solution of (1). Moreover, assume that
(i) For every T and N , there is a constant D depending only on T and N
such that for all |x| , |y| N and all 0 t T ,
|a (t, x) a (t, y)| + | (t, x) (t, y)| D |x y|
(ii) Coecients satisfy the linear growth condition
|a (t, x)| + | (t, x)| D (1 + |x|)
 
2
(iii) E |x| < ,
2
then Xm
(t) converges to X(t) in L .
2624 F. Mohammadi MJOM

Proof. Consider the functions z1 (t) = a (t, X(t)) , z2 (t) = (t, X(t)) and their
second kind Chebyshev wavelet approximated, i.e.,
z1 (t) = a (s, X(s)) , z2 (t) = (s, X(s)) , (25)

z1 (t) = a
(s, Xm
(s)) , z2 (t) =
(s, Xm
(s)) , (26)
and
z1m

(t) = a (s, Xm
(s)) , z2m

(t) = (s, Xm
(s)) . (27)

(t) = X(t) Xm
The error function em (t) can be written as

t t
(t) =
em (z1 (s) z1 (s))ds + (z2 (s) z2 (s)) dB(s), (28)
0 0

as a result we get
 2  t 2
t   
   
(z1 (s) z1 (s))ds +E  (z1 (s) z1 (s))dB(s)
 
2
E em (t) = 2 E   ,
   
0 0

(29)
from the Ito isometry we have
t
 t
2 E (z1 (s) z1 (s)) 2 ds+ E (z2 (s) z2 (s)) 2 ds ,
E em
(t) = 2
0 0
(30)

t t
 2  2
8 E  z1 (s)z1m

(s)  ds + E  z1m (s) z1 (s)  ds
0 0
(31)

t t
 2  2
+8 E  z2 (s)z2m

(s)  ds+ E  z2m (s) z2 (s)  ds .
0 0
(32)
Theorem 3.3 shows that second kind Chebyshev wavelet expansion of any
function f converges uniformly to f . So, for any there exists n such that
 2
E  zim (s) zi (s)  , i = 1, 2, (33)
so we get
2
E em
(t)

t t
 2  m 2
+ 8 E  z1 (s) z1m

(s)  ds+ E  z2 (s) z2 (s)  ds, (34)
0 0
Vol. 13 (2016) Second Kind Chebyshev Wavelet 2625

from Lipschitz condition we can write


t
2 2 2
E em
(t) + 16D em
(s) (35)
0

The Gronwall inequality for this inequality results


2
E em
(t) 0, (36)
2
(t) converges to X(t) in Lw .
and so Xm 

7. Numerical Examples
In this section, we demonstrate the eciency of the proposed second kind
Chebyshev wavelet method (SKCHWM) in Sect. 5 with some illustrative
examples. It will be shown that the numerical results derived by the proposed
method are comparable to exact solution and those of other existing methods.
From now on N is the number of iterations. Moreover, let X(t) denote the

exact solution and X(t) be the second kind Chebyshev wavelets solution
computed by the presented method. The error of the approximate solution is
dened as
i ) X(ti ) .
E(ti ) = M ax X(t (37)

Example 1. Consider the nonlinear stochastic It


o-Volterra integral equation
(population growth problem) [9,11]
t t
1
X(t) = + X(s) (1 X(s))ds + X(s)dB(s), s, t [0, 1] , (38)
2
0 0

where X(t) is an unknown stochastic process dened on the probability space


(, , P ), and B(t) is a Brownian motion process. The exact solution of this
stochastic Ito-Volterra integral equation is

e0.5t+B(t)
X(t) = t . (39)
2 + 0 e0.5s+B(s) ds

The stochastic operational matrix of second kind Chebyshev wavelet and


the presented method in Sect. 5 is employed for deriving numerical solution
of this It
o-Volterra integral equation. The exact and approximate solution
computed by presented method for m = 128 and N = 100 is represented
in Fig. 1. To verify the proposed SKCHWM, the obtained numerical results
have been compared with previously published numerical ones of Ref. [11].
Table 1 shows the error means of the presented SKCHWM and the Bernstein
polynomials method [11] for some points in the interval [0, 1].
2626 F. Mohammadi MJOM

Figure 1. The approximate solution and exact solution for


m
= 128 and N = 100

Table 1. The error means of Example 1 for m


= 128 and
N = 100

t SKCHWM Ref. [11]


0.1 0.007191 0.021256
0.2 0.008549 0.044276
0.3 0.031913 0.070860
0.4 0.053830 0.096359
0.5 0.083752 0.117920
0.6 0.094932 0.141850
0.7 0.084362 0.162259
0.8 0.083965 0.188730
0.9 0.191201 0.209053

Example 2. Let us consider the following stochastic Ito-Volterra integral


equation[4,7,9]
t t
1
X(t) = + cos(s)X(s)ds + sin(s)X(s)dB(s), s, t [0, 1] , (40)
12
0 0

where X(t) is an unknown stochastic process dened on the probability space


(, , P ), and B(t) is a Brownian motion process. The exact solution of this
stochastic Volterra integral equation is

t
1 t sin(2t)
X(t) = exp + sin(t) + + sin(s)dB(s). (41)
12 4 8
0
Vol. 13 (2016) Second Kind Chebyshev Wavelet 2627

Figure 2. The approximate solution and exact solution for


m
= 128
Table 2. Numerical results for m
= 128 and N = 100

t Exact Ref. [7] CHWM SKCHWM


0.1 0.092646 0.092748 0.092760 0.092763
0.2 0.098622 0.098756 0.098717 0.098715
0.3 0.104411 0.103398 0.104510 0.104494
0.4 0.116055 0.116955 0.116577 0.116523
0.5 0.128423 0.130957 0.129383 0.129568

This stochastic Ito-Volterra integral equation has been solved numerically


using BPFs [4], hat basis functions [7] and Chebyshev wavelet [9]. Here,
the stochastic operational matrix of second kind Chebyshev wavelet and the
proposed method in Sect. 5 are employed to approximate solution of this
stochastic Ito-Volterra integral equation. The exact and approximate solution
computed by the presented SKCHWM for m = 128 and N = 100 is plotted
in Fig. 2. Moreover, Table 2 shows exact solution and numerical solution
obtained by the SKCHWM, generalized hat basis functions [7] and Chebyshev
wavelet method (CHWM) [9] for some points in the interval [0, 1]. As these
numerical results show, the SKCHWM is ecient and accurate for solving
stochastic Ito-Volterra integral equation.

Example 3. Consider the following stochastic It


o-Volterra integral equation
[4,7,9]
t t
X(t) = 1 + s2 X(s)ds + sX(s)dB(s), s, t [0, 1] , (42)
0 0
where X(t) is an unknown stochastic process dened on the probability space
(, , P ), and B(t) is a Brownian motion process. The exact solution of this
2628 F. Mohammadi MJOM

Figure 3. The approximate solution and exact solution for


m
= 128

Table 3. Numerical results for m


= 128 and N = 100

t Exact Ref. [7] CHWM SKCHWM


0.1 1.006472 1.007581 1.006151 1.006231
0.2 0.972661 0.973908 0.972836 0.972194
0.3 0.940384 0.931163 0.940423 0.940572
0.4 0.963411 0.961170 0.962597 0.962178
0.5 0.994197 1.014848 0.994243 0.994263

stochastic Volterra integral equation is



3 t
t
X(t) = exp + sdB(s) . (43)
6
0

The proposed SKCHWM in Sect. 5 are used for solving this stochastic Ito-
Volterra integral equation. The exact and approximate solution computed
by the presented method for m = 128 and N = 100 is shown in Fig. 3.
Moreover, Table 3 shows the exact and numerical solution derived by the pre-
sented method, the generalized hat basis functions [7] and Chebyshev wavelet
method (CHWM) [9] for some points in the interval [0, 1]. The numerical
results conrm eciency and accuracy of the proposed SKCHWM.

8. Conclusion
A new stochastic operational matrix for the second kind Chebyshev wavelets
is presented and a general procedure for forming this matrix is given. A
Vol. 13 (2016) Second Kind Chebyshev Wavelet 2629

wavelet Galerkin method based on this stochastic operational matrix is pro-


posed for solving stochastic It
o-Voltera integral equations. Convergence and
error analysis of the second kind Chebyshev wavelets basis are investigated.
To reveal the accuracy and eciency of the proposed method some numer-
ical examples are included. The numerical results are compared with exact
solution and those of other existing methods.

References
[1] Kloeden, P.E., Platen, E.: Numerical Solution of Stochastic Dierential Equa-
tions. Springer-Verlag, Berlin (1999)
[2] Oksendal, B. : Stochastic Dierential Equations: An Introduction with Appli-
cations. Springer, New York (1998)
[3] Higham, D.J.: An algorithmic introduction to numerical simulation of stochas-
tic dierential equations. SIAM Rev. 43(3), 525546 (2001)
[4] Maleknejad, K., Khodabin, M., Rostami, M.: Numerical solution of stochastic
Volterra integral equations by a stochastic operational matrix based on block
pulse functions. Math. Comput. Model 55, 791800 (2012)
[5] Khodabin, M., Maleknejad, K., Rostami, M., Nouri, M.: Numerical approach
for solving stochastic VolterraFredholm integral equations by stochastic oper-
ational matrix. Comput. Math. Appl. 64, 19031913 (2012)
[6] Khodabin, M., Maleknejad, K., Hosseini, F.: Shekarabi, application of tri-
angular functions to numerical solution of stochastic volterra integral equa-
tions. IAENG Int. J. Appl. Math. 43(1), 19 (2011)
[7] Heydari, M.H., Hooshmandasl, M.R., Maalek, F.M., Cattani, C.: A computa-
tional method for solving stochastic It o-Volterra integral equations based on
stochastic operational matrix for generalized hat basis functions. J. Comput.
Phys 270, 402415 (2014)
[8] Heydari, M.H., Hooshmandasl, M.R., Cattani, C., Ghaini, F.M.M.: An e-
cient computational method for solving nonlinear stochastic It o integral equa-
tions: application for stochastic problems in physics. J. Comput. Phys 283, 148
168 (2015)
[9] Mohammadi, F.: A wavelet-based computational method for solving stochastic
It
o-Volterra integral equations. J. Comput. Phys 298, 254265 (2015)
[10] Mohammadi, F.: A Chebyshev wavelet operational method for solving stochas-
tic Volterra-Fredholm integral equations. Int. J. Appl. Math. Res. 4(2), 217
227 (2015)
[11] Asgari, M., Hashemizadeh, E., Khodabin, M., Maleknejad, K.: Numeri-
cal solution of nonlinear stochastic integral equation by stochastic opera-
tional matrix based on Bernstein polynomials. Bull. Math. Soc. Sci. Math.
Roumanie 57(1), 312 (2014)
[12] Murge, M.G., Pachpatte, B.G.: Successive approximations for solutions of sec-
ond order stochastic integro-dierential equations of It o type. Indian. J. Pure.
Appl. Math. 21(3), 260274 (1990)
[13] Cortes, J.C., Jodar, L., Villafuerte, L.: Numerical solution of random dier-
ential equations: a mean square approach. Math. Comput. Model 45, 757
765 (2007)
2630 F. Mohammadi MJOM

[14] Jankovic, S., Ilic, D.: One linear analytic approximation for stochastic integro-
dierential equations. Acta Math. Sci. 30, 10731085 (2010)
[15] Strang, G.: Wavelets and dilation equations: a brief introduction. SIAM
Rev. 31(4), 614627 (1989)
[16] Boggess, A., Narcowich, F.J.: A First Course in Wavelets with Fourier Analysis.
John Wiley & Sons (2009)
[17] Beylkin, G., Coifman, R., Rokhlin, V.: Fast wavelet transforms and numerical
algorithms I. Commun. Pure Appl. Math. 44, 141183 (1991)
[18] Razzaghi, M., Youse, S.: The Legendre wavelets operational matrix of inte-
gration. Int. J. Syst. Sci. 32(4), 495502 (2001)
[19] Lepik, U.: Numerical solution of dierential equations using Haar
wavelets. Math. Comput. Simulat. 68, 127143 (2005)
[20] Babolian, E., Fattahzadeh, F.: Numerical solution of dierential equations by
using Chebyshev wavelet operational matrix of integration. Appl. Math. Com-
put. 188(1), 417426 (2007)
[21] Li, Y.: Solving a nonlinear fractional dierential equation using Chebyshev
wavelets. Commun. Nonlinear Sci. Numer. Simul. 15(9), 22842292 (2010)
[22] Adibi, H., Assari, P.: Chebyshev wavelet method for numerical solution of
Fredholm integral equations of the rst kind, Math. Probl. Eng. 2010 (2010)
[23] Liu, N., Lin, E.B.: Legendre wavelet method for numerical solutions of partial
dierential equations. Numer. Meth. Part. Dier. Equ. 26(1), 8194 (2010)
[24] Mohammadi, F., Hosseini, M.M.: Syed Tauseef Mohyud-Din, Legendre wavelet
Galerkin method for solving ordinary dierential equations with non-analytic
solution. Int. J. Syst. Sci. 42(2), 579585 (2011)
[25] Mohammadi, F.: Numerical solution of BagleyTorvik equation using Cheby-
shev wavelet operational matrix of fractional derivative. Int. J. Adv. Appl.
Math. Mech. 2(1), 8391 (2014)
[26] Mohammadi, F.: An extended Legendre wavelet method for solving dierential
equations with non analytic solutions. J. Math. Exten. 8, 5574 (2015)
[27] Gupta, A.K., Ray, S.S.: Numerical treatment for the solution of fractional
fth-order SawadaKotera equation using second kind Chebyshev wavelet
method. Appl. Math. Model 39(17), 51215130 (2015)
[28] Wang, Y., Fan, Q.: The second kind Chebyshev wavelet method for solving frac-
tional dierential equations. Appl. Math. Comput. 218(17), 85928601 (2012)
[29] Zhu, L., Fan, Q.: Solving fractional nonlinear Fredholm integro-dierential
equations by the second kind Chebyshev wavelet. Commun. Nonlinear Sci.
Numer. Simul. 17(6), 23332341 (2012)
[30] Jiang, Z., Schoufelberger, W., Thoma, M., Wyner, A.: Block pulse functions
and their applications in control systems. Springer-Verlag, New York,
Inc. (1992)
[31] Mason, J.C., Handscomb, D.C.: Chebyshev polynomials. CRC Press (2002)
Vol. 13 (2016) Second Kind Chebyshev Wavelet 2631

Fakhrodin Mohammadi
Department of Mathematics
Hormozgan University
P. O. Box 3995
Bandarabbas, Iran
e-mail: f.mohammadi62@hotmail.com

Received: July 15, 2015.


Revised: September 7, 2015.
Accepted: October 1, 2015.

Anda mungkin juga menyukai