JAM Journal of Applied Mathematics 1687-0042 1110-757X Hindawi Publishing Corporation 10.1155/2014/134272 134272 Research Article Bernstein Collocation Method for Solving Nonlinear Fredholm-Volterra Integrodifferential Equations in the Most General Form Akyüz-Daşcıoğlu Ayşegül 1 İşler Acar Neşe 2 Güler Coşkun 3 Shahzad Naseer 1 Department of Mathematics Faculty of Arts & Sciences, Pamukkale University, 20070 Denizli Turkey pau.edu.tr 2 Department of Mathematics Faculty of Arts & Sciences, Mehmet Akif Ersoy University, 15030 Burdur Turkey mehmetakif.edu.tr 3 Department of Mathematical Engineering Faculty of Chemistry and Metallurgy, Yildiz Technical University, 34210 Istanbul Turkey yildiz.edu.tr 2014 1082014 2014 14 05 2014 17 07 2014 10 8 2014 2014 Copyright © 2014 Ayşegül Akyüz-Daşcıoğlu et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

A collocation method based on the Bernstein polynomials defined on the interval [a,b] is developed for approximate solutions of the Fredholm-Volterra integrodifferential equation (FVIDE) in the most general form. This method is reduced to linear FVIDE via the collocation points and quasilinearization technique. Some numerical examples are also given to demonstrate the applicability, accuracy, and efficiency of the proposed method.

1. Introduction

The quasilinearization method was introduced by Bellman and Kalaba  to solve nonlinear ordinary or partial differential equations as a generalization of the Newton-Raphson method. The origin of this method lies in the theory of dynamic programming. In this method, the nonlinear equations are expressed as a sequence of linear equations and these equations are solved recursively. The main advantage of this method is that it converges monotonically and quadratically to the exact solution of the original equations . Therefore, the quasilinearization method is an effective approach for obtaining approximate solutions of nonlinear equations such as differential equations , functional equations [8, 9], integral equations , and integrodifferential equations .

In this paper, we consider the nonlinear FVIDE in the general form (1)g(x,y(x),y(x),,y(m)(x))=λ1abf(x,t,y(t),y(t),,y(m)(t))dt+λ2axv(x,t,y(t),y(t),,y(m)(t))dt, under the initial (2)k=0m-1τjky(k)(c)=μj;j=0,1,,m-1,c[a,b], or boundary conditions (3)k=0m-1[αjky(k)(a)+βjky(k)(b)]=γj;j=0,1,,m-1. Here g:[a,b]×Rm+1R, f:[a,b]×[a,b]×Rm+1R, and v:[a,b]×[a,b]×Rm+1R are known functions, αjk, βjk, τjk, μj, γj, λ1, and λ2 are known constants, and y(x) is an unknown function.

Besides, we approximate to the nonlinear FVIDE (1) by the generalized Bernstein polynomials defined on the interval [a,b] as (4)y(x)Bn(y;x)=i=0ny(a+(b-a)in)pi,n(x), where pi,n(x) denotes the generalized Bernstein basis polynomials of the form (5)pi,n(x)=1(b-a)n(ni)(x-a)i(b-x)n-i;i=0,1,,n. For convenience, we set pi,n(x)=0, if i<0 or i>n.

Bernstein polynomials have many useful properties such as the positivity, continuity, differentiability, integrability, recursion’s relation, symmetry, and unity partition of the basis set over the interval [a,b]. For more information about the Bernstein polynomials, see [16, 17]. Recently, these polynomials have been used for the numerical solutions of differential equations [4, 18, 19], integral equations , and integrodifferential equations .

Now, we give two main theorems for the generalized Berntein polynomials and their basis forms that were proved by Akyuz Dascioglu and Isler  as follows.

Theorem 1.

If yCk[a,b], for some integer m0, then (6)limnBn(k)(y;x)=y(k)(x);k=0,1,,m converges uniformly.

Proof.

The above theorem can be easily proved by applying transformation t=(x-a)/(b-a) to the theorem given on the interval [0,1] by Phillips .

Theorem 2 (see [<xref ref-type="bibr" rid="B5">4</xref>]).

There is a relation between generalized Bernstein basis polynomials matrix and their derivatives in the form (7)P(k)(x)=P(x)Nk;k=1,,m such that (8)P(x)=[p0,n(x)p1,n(x)pn,n(x)]. Here the elements of (n+1)×(n+1) matrix N=(dij), i,j=0,1,,n, are defined by (9)dij=1b-a{n-i;ifj=i+12i-n;ifj=i-i;ifj=i-10;otherwise.

In highlight of these theorems, a collocation method based on the generalized Bernstein polynomials, given in Section 2, is developed for the approximate solutions of the nonlinear FVIDE in the most general form (1) via the quasilinearization technique iteratively. In Section 3, some numerical examples are presented for exhibiting the accuracy and applicability of the proposed method. Finally, the paper ends with the conclusions in Section 4.

2. Method of the Solution

Our aim is to obtain a numerical solution of the nonlinear FVIDE in the general form (1) under conditions (2) or (3) in terms of the generalized Bernstein polynomials. For this, we firstly express this nonlinear equation as a sequence of linear FVIDEs via the quasilinearization technique iteratively. After that, using the collocation points yields the system of linear algebraic equations. This system represents a matrix equation given by the following theorem. Finally, solving this system with the conditions we get the desired approximate solution.

Theorem 3.

Let xs be collocation points defined on the interval [a,b], and let the functions g, f, and v be able to expand by Taylor series with respect to y(k);k=0,1,,m. Suppose that nonlinear FVIDE (1) has the generalized Bernstein polynomial solution. Then, the following matrix relation holds: (10)[k=0m(Gr,kP-λ1Fr,k-λ2Vr,k)Nk]Yr+1=Hr;r=0,1,, where N is defined in Theorem 2, P=[pi,n(xs)], Gr,k=diag[Gr,k(xs)], Fr,k=[Fs,ir,k], and Vr,k=[Vs,ir,k] are (n+1)  ×(n+1) matrices, and Yr+1=[yr+1(a+(b-a)i/n)] and Hr=[hr(xs)] are (n+1)×1 matrices for i,s=0,1,,n, such that (11)Gr,k(xs)=gyr(k)(xs,yr(xs),yr(xs),,yr(m)(xs)),Fs,ir,k=abfyr(k)(xs,t,yr(t),yr(t),,yr(m)(t))pi,n(t)dt,Vs,ir,k=axsvyr(k)(xs,t,yr(t),yr(t),,yr(m)(t))pi,n(t)dt. Here r is iteration index, pi,n(x) is generalized Bernstein basis polynomials, and hr(x) is given in the following proof.

Proof.

Firstly, by applying the quasilinearization method to the nonlinear FVIDE (1), we obtain a sequence of linear FVIDEs: (12)g(x,yr(x),yr(x),,yr(m)(x))+k=0m(yr+1(k)(x)-yr(k)(x))gyr(k)=k=0m[λ1ab(yr+1(k)(t)-yr(k)(t))fyr(k)dtmmmv+λ2ax(yr+1(k)(t)-yr(k)(t))vyr(k)dt]+λ1abf(x,t,yr(t),yr(t),,yr(m)(t))dt+λ2axv(x,t,yr(t),yr(t),,yr(m)(t))dt, where the expressions gyr(k), fyr(k), and vyr(k) represent partial differentiation of the functions g, f, and v with respect to yr(k), and these are defined, respectively, as (13)gyr(k)=gyr(k)(x,yr(x),yr(x),,yr(m)(x)),fyr(k)=fyr(k)(x,t,yr(t),yr(t),,yr(m)(t)),vyr(k)=vyr(k)(x,t,yr(t),yr(t),,yr(m)(t)). Here y0(x) is a reasonable initial approximation of the function y(x), and yr(x) is always considered known and is obtained from previous iteration. The recurrence relation (12) can now be written compactly in the form (14)k=0m[gyr(k)yr+1(k)(x)-λ1abfyr(k)yr+1(k)(t)dtmmv-λ2axvyr(k)yr+1(k)(t)dt]=hr(x) denoting hr(x): (15)hr(x)=-g(x,yr,yr,,yr(m))+k=0m[gyr(k)yr(k)(x)-λ1abfyr(k)yr(k)(t)dtmmmmm-λ2axvyr(k)yr(k)(t)dt]+λ1abf(x,t,yr(t),yr(t),,yr(m)(t))dt+λ2axv(x,t,yr(t),yr(t),,yr(m)(t))dt. Notice that (14) is a linear FVIDE with variable coefficients, since yr(x) is known function of x. yr+1(x) is an unknown function that has the Bernstein polynomial solution; also this function and its derivatives can be expressed by (16)yr+1(k)(x)Bn(k)(yr+1;x)=P(k)(x)Yr+1;r=0,1,. By utilizing Theorem 2 and collocation points, above relation becomes (17)yr+1(k)(xs)=P(xs)NkYr+1;k=0,1,,m. Substituting the collocation points and the relation (17) into (14), we obtain a linear algebraic system: (18)k=0m[Gr,k(xs)P(xs)-λ1Fr,k(xs)-λ2Vr,k(xs)]NkYr+1=hr(xs). Here Fr,k(xs) and Vr,k(xs) are denoted by (19)Fr,k(xs)=abfyr(k)(xs,t,yr(t),yr(t),,yr(m)(t))P(t)dt,Vr,k(xs)=axsvyr(k)(xs,t,yr(t),yr(t),,yr(m)(t))P(t)dt.P(t) and Gr,k(x) are defined, respectively, in Theorems 2 and 3.

For s=0,1,,n, the system (18) can be written compactly in the matrix form (20)WrYr+1=Hr;r=0,1, so that (21)Wr=k=0m(Gr,kP-λ1Fr,k-λ2Vr,k)Nk, where matrices are clearly (22)Gr,k=[Gr,k(x0)000Gr,k(x1)000Gr,k(xn)],P=[P(x0)P(x1)P(xn)],Fr,k=[Fr,k(x0)Fr,k(x1)Fr,k(xn)],Vr,k=[Vr,k(x0)Vr,k(x1)Vr,k(xn)],Hr=[hr(x0)hr(x1)hr(xn)]. Hence, the proof is completed.

Corollary 4.

For λ1=λ2=0, the nonlinear FVIDE (1) is reduced to the mth order nonlinear differential equation (23)g(x,y(x),y(x),,y(m)(x))=0, and by utilizing Theorem 3, this equation can be written as (24)k=0mGr,kPNkYr+1=H^r;r=0,1,. Here the matrices P, N, and Gr,k are defined as above, and elements of the matrix H^r=[h^r(xs)] are denoted by (25)h^r(xs)=k=0mGr,k(xs)yr(k)(xs)vvvvv-g(xs,yr(xs),yr(xs),,yr(m)(xs)).

Corollary 5.

For m=0, the nonlinear FVIDE (1) is reduced to the nonlinear Fredholm-Volterra integral equation in the form (26)g(x,y(x))=λ1abf(x,t,y(t))dt+λ2axv(x,t,y(t))dt such that y(0)(t)=y(t). From Theorem 3, this equation has the iteration matrix form as (27)[Gr-λ1Fr-λ2Vr]Yr+1=Hr;r=0,1,. Here Gr=diag[Gr(xs)], Fr=[Fr,s,i], Vr=[Vr,s,i], Hr=[hr(xs)], and elements of these matrices are denoted as follows: (28)Gr(xs)=gyr(xs,yr(xs)),Fr,s,i=abfyr(xs,t,yr(t))pi,n(t)dt,Vr,s,i=axsvyr(xs,t,yr(t))pi,n(t)dt,hr(xs)=gyr(xs,yr(xs))yr(xs)-g(xs,yr(xs))-λ1ab[f(xs,t,yr(t))-fyr(xs,t,yr(t))yr(t)]dt-λ2axs[v(xs,t,yr(t))-vyr(xs,t,yr(t))yr(t)]dt.

Now we can solve the nonlinear FVDIE (1) under the initial (2) or boundary (3) conditions as follows.

Step 1.

Firstly, we use Theorem 3 for the nonlinear FVDIE (1) and determine the matrices in (10). This matrix equation is a system of linear algebraic equations with n-unknown coefficients yr+1(a+(b-a)i/n). Let the augmented matrix corresponding equation (10) be denoted by [Wr;Hr].

Step 2.

We need to choose the first iteration function y0(x) for calculating the Wr and Hr. Notice that this function can be obtained in a variety of ways. For instance, it can be obtained from the physical situation for engineering problems. However, a very rough choice for the first iteration function such as initial value is enough for the procedure to converge. We can also consider that the first iteration function as the highest degree polynomial satisfied the given conditions (2) or (3).

Step 3.

From expression (17), initial (2) and boundary (3) conditions can be written in the matrix forms, respectively, (29)IjYr+1=μj,BjYr+1=γj, where the matrices are (30)Ij=k=0m-1τjkP(c)Nk,Bj=k=0m-1[αjkP(a)Nk+βjkP(b)Nk]. Besides, (29) can be denoted by the augmented matrices [Ij;μj] and [Bj;γj].

Step 4.

To obtain the solution of nonlinear FVIDE (1) under the given conditions, we insert the elements of the row matrices [Ij;μj] or [Bj;γj] to the end of the augmented matrix [Wr;Hr]. In this way, we have the new augmented matrix [W~r;H~r], that is, (n+m+1)×(n+1) rectangular matrix.

Step 5.

If rank(W~r)=rank[W~r;H~r]=n+1, then unknown coefficients yr+1 are uniquely determined for each iteration r. This kind of systems can be solved by the Gauss Elimination, Generalized Inverse, and QR factorization methods.

3. Numerical Results

Four numerical examples are given to illustrate the applicability, accuracy, and efficiency of the proposed method. All results are computed by using an algorithm written in Matlab 7.1. Besides, in the tables, the absolute and L2-norm errors are computed numerically on the collocation points xs=a+(b-a)s/n;s=0,1,,n, by the folowing formulas: (31)Eabs=|y(xs)-Bn(yr;xs)|,E2=0xs(y(x)-Bn(yr;x))2dx such that Bn(yr;x) is a Bernstein approximation for the rth iteration function yr and y(x) is an exact solution.

Example 1.

Consider the nonlinear Volterra integrodifferential equation (32)y(x)=2x-12sin(x4)+0xx2tcos(x2y(t))dt;0<x,t<1 under the initial condition y(0)=0. Exact solution of this equation is y(x)=x2. Here g(x,y(x))=y(x)-2x+(1/2)sin(x4) and v(x,t,y(t))=x2tcos(x2y(t)).

Let y0(x)=0 be the first iteration function. From Theorem 3, matrix relation of the above problem can be written as (33)(Gr,1PN+Vr,0)Yr+1=Hr;r=0,1,, where Hr=[hr(xs)], Vr,0=[Vs,ir,0], P=[pi,n(xs)], and Gr,1=In+1 (In+1 is an identity matrix with dimensional (n+1)), because of the gyr=1. Elements of the matrices become (34)hr(xs)=2xs-12sin(xs4)+0xs[xs2tcos(xs2yr(t))+xs4tsin(xs2yr(t))yr(t)]dt,Vs,ir,0=0xsxs4tsin(xs2yr(t))pi,n(t)dt.

A numerical comparison of the proposed method with the Hybrid method  is given in Table 1. The obtained numerical results for two different initial approximations are also listed in Table 1. It can be seen that the computational results of the proposed method are better and more effective for smaller values n and iterations r than the other method, and the choice of the higher degree polynomial for the initial approximation leads to better results.

Comparison of the L2-norm errors for Example 1.

n Presented method Hybrid method 
y 0 ( x ) = 0 y 0 ( x ) = x
r = 2 r = 3 r = 2 r = 3 n = 9 8.8 e - 012
2 7.0 e - 006 1.7 e - 012 8.7 e - 008 3.5 e - 017 n = 17 3.4 e - 012
3 2.2 e - 007 3.8 e - 017 2.2 e - 009 7.2 e - 017 n = 19 2.1 e - 013
4 6.0 e - 008 1.3 e - 016 1.1 e - 009 2.4 e - 017 n = 39 9.7 e - 017

Example 2.

Consider the nonlinear Fredholm-Volterra integrodifferential equation (35)y(x)=-y(x)+2x+x2+x510-132+1401ty3(t)dt-120xy2(t)dt;x,t[0,1] with the initial condition y(0)=0 that exact solution is y(x)=x2. Here the required functions and constants are denoted by g(x,y(x),y(x))=y(x)+y(x)-2x-x2-(x5/10)+(1/32), f(x,t,y(t))=ty3(t), v(x,t,y(t))=y2(t), λ1=1/4, and λ2=-1/2.

We have two choices satisfying the initial condition for the first iteration functions such that y0(x)=0 and y0(x)=x. Let the first iteration function be y0(x)=x, because of the higher degree. From Theorem 3, matrix relation of the above problem is (36)(Gr,0P+Gr,1PN-14Fr,0+12Vr,0)Yr+1=Hr;r=0,1,, where Hr=[hr(xs)], Fr,0=[Fs,ir,0], Vr,0=[Vs,ir,0], P=[pi,n(xs)], and Gr,0=Gr,1=In+1 because of the gyr=gyr=1. Elements of these matrices are as follows: (37)Fs,ir,0=013tyr2(t)pi,n(t)dt,Vs,ir,0=0xs2yr(t)pi,n(t)dt,hr(xs)=2xs+xs2+xs510-132+14×01[-2tyr3(t)]dt-120xs[-yr2(t)]dt.

Table 2 contains a numerical comparison of the proposed method between the numerical method based on fixed point theorem  and direct method by using triangular functions . The table reveals that convergence of the presented method is faster and more accurate than the others.

Comparison of the absolute errors for Example 2.

x Presented method Direct method  Fixed point method 
n = 2 , r=1 n = 2 , r=2 n = 2 , r=3 n = 2 , r=4 n = 15 n = 4
0.1 5.7 e - 004 1.8 e - 008 2.5 e - 017 6.1 e - 018 9.8 e - 004 2.9 e - 005
0.2 8.7 e - 003 2.9 e - 008 3.5 e - 017 9.0 e - 018 7.0 e - 004 1.2 e - 005
0.3 1.0 e - 003 3.6 e - 008 4.2 e - 017 1.1 e - 017 7.4 e - 004 9.0 e - 005
0.4 1.1 e - 003 4.0 e - 008 4.6 e - 017 1.2 e - 017 1.1 e - 003 3.1 e - 004
0.5 9.3 e - 004 4.1 e - 008 4.7 e - 017 1.2 e - 017 1.6 e - 004 1.1 e - 004
0.6 6.7 e - 004 3.8 e - 008 4.4 e - 017 1.2 e - 017 1.1 e - 003 1.0 e - 003
0.7 2.5 e - 004 3.2 e - 008 3.8 e - 017 1.0 e - 017 8.2 e - 004 2.0 e - 003
0.8 3.0 e - 004 2.2 e - 008 2.9 e - 017 7.6 e - 018 8.2 e - 004 2.9 e - 003
0.9 1.0 e - 003 8.4 e - 009 1.6 e - 017 4.2 e - 018 1.1 e - 003 1.1 e - 003
1.0 1.8 e - 003 8.4 e - 009 0 0 1.5 e - 004 1.8 e - 003

Example 3.

Consider the third-order nonlinear Fredholm integrodifferential equation (38)y′′′(x)=-ex+-11ex-2ty2(t)dt with the boundary conditions (39)y(0)=y(0)=1,y(1)=e. The exact solution of the above equation is y(x)=ex. Here g(x,y′′′(x))=y′′′(x)+ex and f(x,t,y(t))=ex-2ty2(t).

Let the first iteration function be y0(x)=1+x+(e-2)x2. From Theorem 3, matrix relation of the above problem can be denoted by (40)(Gr,3PN3-Fr,0)Yr+1=Hr;r=0,1, such that Hr=[hr(xs)], Fr,0=[Fs,ir,0], P=[pi,n(xs)], and Gr,3=In+1. Here elements of matrices are, respectively, (41)Fs,ir,0=-112exs-2tyr(t)pi,n(t)dt,hr(xs)=-exs--11exs-2tyr2(t)dt.

In Table 3, absolute errors of the proposed method are given for different values n and iterations r. The table shows that the presented method converges quite rapidly for increasing values n and iterations r. Besides, the absolute error of the homotopy analysis method  given with figure is approximately 2.0e-006 for iteration r=6. Therefore, we can say that the proposed method has more effective numerical results than the other methods.

Absolute errors for Example 3.

x n = 6 n = 8
r = 1 r = 3 r = 5 r = 1 r = 3 r = 5
0 . 01 2.6 e - 004 2.6 e - 004 2.6 e - 004 5.1 e - 009 1.4 e - 008 1.4 e - 008
0.02 2.5 e - 004 2.6 e - 004 2.6 e - 004 2.2 e - 008 1.4 e - 008 1.4 e - 008
0.03    2.5 e - 004 2.5 e - 004 2.5 e - 004 6.6 e - 008 1.4 e - 008 1.4 e - 008
0.04 2.5 e - 004 2.5 e - 004 2.5 e - 004 1.3 e - 007 1.4 e - 008 1.4 e - 008
0.05 2.4 e - 004 2.5 e - 004 2.5 e - 004 2.0 e - 007 1.4 e - 008 1.4 e - 008
0.06 2.4 e - 004 2.4 e - 004 2.4 e - 004 3.0 e - 007 1.4 e - 008 1.4 e - 008
0.07 2.4 e - 004 2.4 e - 004 2.4 e - 004 4.1 e - 007 1.4 e - 008 1.4 e - 008
0.08 2.3 e - 004 2.4 e - 004 2.4 e - 004 5.3 e - 007 1.4 e - 008 1.4 e - 008
0.09 2.3 e - 004 2.3 e - 004 2.3 e - 004 6.7 e - 007 1.4 e - 008 1.4 e - 008
0.1 2.3 e - 004 2.3 e - 004 2.3 e - 004 8.3 e - 007 1.3 e - 008 1.3 e - 008

x n = 10 n = 12
r = 1 r = 3 r = 5 r = 1 r = 3 r = 5

0.01 9.1 e - 009 8.4 e - 014 8.4 e - 014 9.1 e - 009 1.1 e - 016 1.9 e - 016
0.02 3.6 e - 008 1.2 e - 013 1.2 e - 013 3.6 e - 008 1.9 e - 016 9.6 e - 017
0.03 8.0 e - 008 3.0 e - 013 3.0 e - 013 8.0 e - 008 2.5 e - 016 7.0 e - 016
0.04 1.4 e - 007 3.4 e - 013 3.4 e - 013 1.4 e - 007 5.4 e - 016 4.2 e - 016
0.05 2.2 e - 007 1.3 e - 013 1.3 e - 013 2.2 e - 007 1.9 e - 015 1.5 e - 015
0.06 3.1 e - 007 4.7 e - 013 4.7 e - 013 3.1 e - 007 4.1 e - 015 3.7 e - 015
0.07 4.2 e - 007 1.5 e - 012 1.5 e - 012 4.2 e - 007 7.5 e - 015 7.0 e - 015
0.08 5.5 e - 007 3.2 e - 012 3.2 e - 012 5.5 e - 007 1.2 e - 014 1.2 e - 014
0.09 6.9 e - 007 5.6 e - 012 5.6 e - 012 6.9 e - 007 1.9 e - 014 1.8 e - 014
0.1 8.4 e - 007 8.8 e - 012 8.8 e - 012 8.4 e - 007 2.7 e - 014 2.6 e - 014

Example 4.

Consider the fourth-order Volterra integrodifferential equation (42)y(4)(x)=1+0xe-ty2(t)dt with the boundary conditions (43)y(0)=y(0)=1,y(1)=e,y(1)=e that exact solution is y(x)=ex. Here the functions are g(x,y(4)(x))=y(4)(x)-1 and v(x,t,y(t))=e-ty2(t).

Let the first iteration function be y0(x)=1+x+(2e-5)x2+(3-e)x3. From Theorem 3, matrix form of the above problem can be written as (44)(Gr,4PN4-Vr,0)Yr+1=Hr;r=0,1,, where Hr=[hr(xs)], Vr,0=[Vs,ir,0], P=[pi,n(xs)], and Gr,4=In+1. Here elements of matrices are, respectively, (45)Vs,ir,0=0xs2e-tyr(t)pi,n(t)dt,hr(xs)=1-0xse-tyr2(t)dt.

In Table 4, L2-norm errors of the proposed method are compared with the L2-norm errors computed via the Hybrid of Block-Pulse functions and Lagrange interpolation polynomials  for different values n. While the results of other methods are effective, the computational results of the proposed method are more rapid, effective, for smaller values n.

Comparison of the L2-norm errors for Example 4.

Presented method Hybrid method 
n r = 0 r = 1 r = 2 r = 3 r = 4 n = 15 4.7 e - 008
6 2.0 e - 004 2.1 e - 004 2.1 e - 004 2.1 e - 004 2.1 e - 004 n = 17 3.0 e - 010
8 1.3 e - 008 8.3 e - 009 8.3 e - 009 8.3 e - 009 8.3 e - 009 n = 19 3.7 e - 011
10 1.3 e - 008 1.2 e - 012 1.2 e - 012 1.2 e - 012 1.2 e - 012 n = 31 3.0 e - 009
12 1.4 e - 008 5.4 e - 015 2.0 e - 015 2.0 e - 015 2.0 e - 015 n = 35 4.5 e - 012
15 1.6 e - 008 5.5 e - 015 1.3 e - 014 1.3 e - 014 1.3 e - 014 n = 39 6.0 e - 013
4. Conclusions

In general, nonlinear integrodifferential equations can not be solved analytically. For this reason, numerical solutions of nonlinear equations are needed. With the presented method, we have reduced the nonlinear FVIDE (1) to a sequence of linear equations depending on the collocation points and the iteration function, and then, combining the conditions with obtained linear matrix equation, we have the generalized Bernstein polynomials solution. Besides, the proposed method is valid for both nonlinear differential and integral equations; this is explained in Section 2. If unknown function is continuous on the interval [a,b], then the Bernstein collocation method can be used for solving these kinds of equations via the quasilinearization technique. For convenience, the first iteration function can be choosen as initial value. Besides, this function can be select as the higher degree polynomials satisfied the given conditions for providing the better approximation. These polynomials can be obtained from interpolation or the least square approximation methods. The applicability and accuracy of the proposed method have been tested with some numerical examples. These results have shown that the presented method converges rapidly for all problems. Consequently, all these positive implications lead to applicability of the proposed method for numerical solutions of any other kinds of problems including nonlinear equations.

Conflict of Interests

The authors declare that there is no conflict of interests regarding the publication of this paper.

Acknowledgment

This work is supported by Scientific Research Project Coordination Unit of Pamukkale University with no. 2012FBE036.

Bellman R. E. Kalaba R. E. Quasilinearization and Nonlinear Boundary Value Problems 1965 New York, NY. USA Elsevier MR0178571 Stanley E. L. Quasilinearization and Invariant Imbedding 1968 New York, NY, USA Academic Press Agarwal R. P. Chow Y. M. Iterative methods for a fourth order boundary value problem Journal of Computational and Applied Mathematics 1984 10 2 203 217 10.1016/0377-0427(84)90058-X MR739776 ZBL0541.65055 2-s2.0-5844393936 Akyuz Dascıoglu A. Isler N. Bernstein collocation method for solving nonlinear differential equations Mathematical & Computational Applications 2013 18 3 293 300 MR3113137 Charles A. Baird J. Modified quasilinearization technique for the solution of boundary-value problems for ordinary differential equations Journal of Optimization Theory and Applications 1969 3 4 227 242 10.1007/BF00926525 MR0245207 2-s2.0-3743072102 Mandelzweig V. B. Tabakin F. Quasilinearization approach to nonlinear problems in physics with application to nonlinear ODEs Computer Physics Communications 2001 141 2 268 281 10.1016/S0010-4655(01)00415-5 MR1872036 2-s2.0-0035976505 Ramos J. I. Piecewise quasilinearization techniques for singular boundary-value problems Computer Physics Communications 2004 158 1 12 25 10.1016/j.comphy.2003.11.003 MR2105397 2-s2.0-1142304425 Ahmad B. Ali Khan R. Sivasundaram S. Generalized quasilinearization method for nonlinear functional differential equations Journal of Applied Mathematics and Stochastic Analysis 2003 16 1 33 43 10.1155/S1048953303000030 MR1973077 ZBL1048.34105 2-s2.0-13444288680 Drici Z. McRae F. A. Devi J. V. Quasilinearization for functional differential equations with retardation and anticipation Nonlinear Analysis: Theory, Methods & Applications 2009 70 4 1763 1775 10.1016/j.na.2008.02.079 MR2483597 2-s2.0-57049176587 Caliò F. Munoz F. Marchetti E. Direct and iterative methods for the numerical solution of mixed integral equations Applied Mathematics and Computation 2010 216 12 3739 3746 10.1016/j.amc.2010.05.032 MR2661738 2-s2.0-77953609794 Maleknejad K. Najafi E. Numerical solution of nonlinear Volterra integral equations using the idea of quasilinearization Communications in Nonlinear Science and Numerical Simulation 2011 16 1 93 100 10.1016/j.cnsns.2010.04.002 MR2679163 2-s2.0-77955919642 Pandit S. G. Quadratically converging iterative schemes for nonlinear volterra integral equations and an application Journal of Applied Mathematics and Stochastic Analysis 1997 10 2 169 178 10.1155/S1048953397000208 MR1453469 2-s2.0-53249135133 Ahmad B. A quasilinearization method for a class of integro-differential equations with mixed nonlinearities Nonlinear Analysis: Real World Applications 2006 7 5 997 1004 10.1016/j.nonrwa.2005.09.003 MR2260894 2-s2.0-33745164886 Ahmad B. Khan R. A. Sivasundaram S. Generalized quasilinearization method for integro-differential equations Nonlinear Studies 2001 8 3 331 341 MR1863951 ZBL1001.65143 Wang P. Wu Y. Wiwatanapaphee B. An extension of method of quasilinearization for integro-differential equations International Journal of Pure and Applied Mathematics 2009 54 1 27 37 MR2541710 ZBL1180.65116 2-s2.0-78649796266 Farouki R. T. Rajan V. T. Algorithms for polynomials in Bernstein form Computer Aided Geometric Design 1988 5 1 1 26 10.1016/0167-8396(88)90016-7 MR945302 2-s2.0-0024030404 Lorentz G. G. Bernstein Polynomials 1986 2nd New York, NY, USA Chelsea MR864976 Doha E. H. Bhrawy A. H. Saker M. A. Integrals of Bernstein polynomials: an application for the solution of high even-order differential equations Applied Mathematics Letters 2011 24 4 559 565 10.1016/j.aml.2010.11.013 MR2749745 2-s2.0-78650516844 Işık O. R. Sezer M. Güney Z. A rational approximation based on Bernstein polynomials for high order initial and boundary values problems Applied Mathematics and Computation 2011 217 22 9438 9450 10.1016/j.amc.2011.04.038 MR2804022 2-s2.0-79957481206 Bhattacharya S. Mandal B. N. Use of Bernstein polynomials in numerical solutions of Volterra integral equations Applied Mathematical Sciences 2008 2 33–36 1773 1787 MR2443905 Maleknejad K. Hashemizadeh E. Ezzati R. A new approach to the numerical solution of Volterra integral equations by using Bernstein's approximation Communications in Nonlinear Science and Numerical Simulation 2011 16 2 647 655 10.1016/j.cnsns.2010.05.006 MR2725788 2-s2.0-77956063494 Mandal B. N. Bhattacharya S. Numerical solution of some classes of integral equations using Bernstein polynomials Applied Mathematics and Computation 2007 190 2 1707 1716 10.1016/j.amc.2007.02.058 MR2339763 ZBL1122.65416 2-s2.0-34250632781 Shirin A. Islam M. S. Numerical solutions of Fredholm integral equations using Bernstein polynomials Journal of Scientific Research 2010 2 2 264 272 Singh V. K. Pandey R. K. Singh O. P. New stable numerical solutions of singular integral equations of Abel type by using normalized Bernstein polynomials Applied Mathematical Sciences 2009 3 5–8 241 255 MR2487576 2-s2.0-65549090973 Işık O. R. Sezer M. Güney Z. Bernstein series solution of a class of linear integro-differential equations with weakly singular kernel Applied Mathematics and Computation 2011 217 16 7009 7020 10.1016/j.amc.2011.01.114 MR2775691 2-s2.0-79952450721 Ordokhani Y. Davaei far S. Application of the Bernstein polynomials for solving the nonlinear Fredholm integro-differential equations Journal of Applied Mathematics & Bioinformatics 2011 1 2 13 31 Pandya B. M. Joshi D. C. Solution of a Volterra's population model in a Bernstein polynomial basis Applied Mathematical Sciences 2011 5 69 3403 3410 MR2860548 2-s2.0-84856530177 Phillips G. M. Interpolation and Approximation by Polynomials 2003 New York, NY, USA Springer MR1975918 Marzban H. R. Hoseini S. M. Solution of nonlinear Volterra-Fredholm integrodifferential equations via hybrid of block-pulse functions and Lagrange interpolating polynomials Advances in Numerical Analysis 2012 2012 14 868279 MR3000360 10.1155/2012/868279 Berenguer M. I. Gámez D. López Linares A. J. Fixed point techniques and Schauder bases to approximate the solution of the first order nonlinear mixed Fredholm—Volterra integro-differential equation Journal of Computational and Applied Mathematics 2013 252 52 61 10.1016/j.cam.2012.09.020 MR3054557 2-s2.0-84866785979 Babolian E. Masouri Z. Hatamzadeh-Varmazyar S. Numerical solution of nonlinear Volterra-Fredholm integro-differential equations via direct method using triangular functions Computers & Mathematics with Applications 2009 58 2 239 247 10.1016/j.camwa.2009.03.087 MR2535790 2-s2.0-67349172178 Shidfar A. Molabahrami A. Babaei A. Yazdanian A. A series solution of the nonlinear Volterra and Fredholm integro-differential equations Communications in Nonlinear Science and Numerical Simulation 2010 15 2 205 215 10.1016/j.cnsns.2009.03.015 MR2572561 ZBL1221.65343 2-s2.0-69049118031