MPE Mathematical Problems in Engineering 1563-5147 1024-123X Hindawi Publishing Corporation 498902 10.1155/2013/498902 498902 Research Article Optimal Homotopy Asymptotic Method for Solving Delay Differential Equations Ratib Anakira N. 1 Alomari A. K. 2 Hashim I. 1 Ntouyas Sotiris 1 School of Mathematical Sciences Universiti Kebangsaan Malaysia 43600 Bangi, Selangor Malaysia ukm.my 2 Department of Mathematics Faculty of Science Hashemite University Zarqa 13115 Jordan hu.edu.jo 2013 29 10 2013 2013 04 05 2013 13 09 2013 2013 Copyright © 2013 N. Ratib Anakira 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.

We extend for the first time the applicability of the optimal homotopy asymptotic method (OHAM) to find the algorithm of approximate analytic solution of delay differential equations (DDEs). The analytical solutions for various examples of linear and nonlinear and system of initial value problems of DDEs are obtained successfully by this method. However, this approach does not depend on small or large parameters in comparison to other perturbation methods. This method provides us with a convenient way to control the convergence of approximation series. The results which are obtained revealed that the proposed method is explicit, effective, and easy to use.

1. Introduction

Delay differential equation (DDE) is a form of differential equations in which derivative of the unknown function in a given time t is specified in terms of the values at an earlier point in time.

DDEs have the general form (1)ui(x)=f(x,ui(x),ui(ξj(x))),i=1,2,,M,j=1,2,,N, where ξj(x)=ajx+bj is the delay function.

Many problems of physics, biological models, control system, and medical and biochemical fields are modelled by DDEs. Recent studies in such diverse fields have shown that DDEs play an important role in explaining many different phenomena. Patel et al.  introduced an iterative scheme for the optimal control systems described by DDEs with a quadratic cost functional. In physiology, Glass and Mackey  applied time delays to many physiological models. Busenberg and Tang  created a model for cell cycle by delay equations. In recent years, DDEs are used to design models as HIV-1 therapy for fighting a virus with another virus .

In the last years, a great deal of attention has been devoted to study DDEs. Hence, they are solved by numerical method and approximation approach, such as Adomian decomposition method [5, 6], homotopy perturbation method (HPM) [7, 8], multiquadric approximation scheme [9, 10], variational iteration method (VIM) [8, 11, 12], spline methods , homotopy analysis method (HAM) , Chebyshev polynomials , Galerkin method , Legendre wavelet method , differential transform method , and Runge-Kutta method . Recently, a new approach of homotopy which is called optimal homotopy asymptotic method (OHAM) was proposed and developed by Marinca et al.  for the approximate solution nonlinear problems of thin film flow of a fourth-grade fluid and for the study of the behavior of nonlinear mechanical vibration of electrical machines. In OHAM, the control and adjustment of the convergence region are provided in a convenient way. Furthermore, the OHAM has been built in convergence criteria similar to those of HAM but with greater degree of flexibility. Islam et al.  have applied this method successfully to nonlinear problems and have also shown its effectiveness and accuracy. Idrees et al.  used OHAM to study the squeezing flow between two infinite planar plates slowly approaching each other.

The aim of this paper is to apply OHAM to get an approximate analytic solution of DDEs. The capability of this approach is tested upon several examples which offer an approximate solution in a series form that converges to exact solution in few terms. The rest of this paper is organized as follows. In Section 2, we describe the basic idea of OHAM. In Section 3, we provide the convergent theorem for this type of equations. Section 4 presents several examples to demonstrate the efficiency of the framework. The conclusion of this study is presented in Section 5.

2. Description of the Method

In this section, framework of the proposed method is given and represented in the following differential equation: (2)Li(ui(x))+gi(x)+Ni(ui(x),ui(ξj(x)))=0,iiiiiiiiiiiiiiiiiiiii=1,2,,N,j=1,2,,M, where Li are the linear operators and Ni are the nonlinear operators contain delay function, ui(x) is an unknown function, x denotes an independent variable, gi(x) is a known function, and ξj(x) are the delay functions.

According to OHAM, we construct a homotopy hi(vi(x,p),p):R×[0,1]R which satisfies (3)(1-p)Li[(vi(x,p)-ui,0(x)]=Hi(p)[Liv(x,p)+gi(x)+Ni(vi(x,p),vi(ξj(x),p))], where xR, p[0,1] is an embedding parameter, Hi(p) is a nonzero auxiliary function for p0,H(0)=0, and v(x,p) is an unknown function. Obviously, when p=0 and p=1 it holds that vi(x,0)=ui,0(x) and vi(x,1)=vi(x), respectively. Thus, as p varies from 0 to 1, the solution vi(x,p) approach from ui,0 to vi(x), where ui,0(x) is the initial guess that satisfies the linear operator and the initial conditions (4)Li(ui,0(x))+g(x)=0. Next, we choose the auxiliary function Hi(p) in the form (5)Hi(p)=pC1+p2C2+p3C3+, where C1,C2,C3, are convergence control parameters which can be determined later. H(p) can be expressed in another form as reported by Herişanu and Marinca .

To get an approximate solution, we expand vi(x,p,Ck) in Taylor’s series about p in the following manner: (6)vi(x,p,Ck)=ui,0(x)+k=1ui,k(x,C1,C2,,Ck)pk. By substituting (6) into (3) and equating the coefficient of like powers of p, we obtain the following linear equations. Define the vectors (7)Ci={C1,C2,,Ci},ui,s={ui,0(x),ui,1(x,C1),,ui,s(x,Cs),ui,0(ξj(x)),ui,1(ξj(x),C1),,ui,s(ξj(x),Cs)(x,Cs),ui,0(ξj(x))}, where s=1,2,3, and j=1,2,,M. The zeroth-order problem is given by (4), and the first- and second-order problems are given as (8)Li(ui,1(x))=C1Ni,0(ui,0)+g(x),(9)Li(ui,2(x))-Li(ui,1(x))=C2Ni,0(ui,0)+C1[Li(ui,1(x))+Ni,1(ui,1)]. The general governing equations for uk(x) are (10)L(uk(x))-L(uk-1(x))=CkN0(u0(x))+m=1k-1Cm[L(uk-m(x))+Nk-m(ui,k-1)], where k=2,3, and Ni,m(ui,0(x),ui,1(x),,ui,m(x)) is the coefficient of pm in the expansion of N(v(x,p)) about the embedding parameter p: (11)Ni(v(x,p,Ci))=Ni,0(u0(x))+m=1Ni,m(ui,m)pm. It has been observed that the convergence of the series (6) depends upon the auxiliary constants C1,C2,C3,. If it is convergent at p=1, one has (12)vi(x,Ck)=u0(x)+k=1ui,k(x,C1,C2,,Ck). The result of the mth-order approximation is given as (13)v~i(x,Ck)=u0(x)+k=1mui,k(x,C1,C2,,Ck). Substituting (12) into (2) yields the following residual: (14)Ri(x,C1,C2,C3,,Cm)=Li(v~i(x,C1,C2,C3,,Cm))+gi(x)+Ni(v~i(x,C1,C2,C3,,Cm)). If Ri=0, then u~ will be the exact solution. Generally such a case will not arise for nonlinear problems, but we can minimize the functional (15)Ji(C1,C2,C3,,Cm)=abRi2(x,C1,C2,C3,,Cm)dx, where a and b are the endpoints of the given problem. The unknown convergence control parameters Ci(i=1,2,3,,m) can be calculated from the system of equations (16)JiCi=0,i=1,2,,m. It should be noted that our process included the auxiliary function Hi(p) which provides us an easy way to set and optimally control the convergent area and the rate of the solution series.

3. Convergence Theorem

In this section, we introduce the convergence of the solution for DDEs.

Theorem 1.

If the series (12) converges to u(x), where uk(x)L(R+) is produced by (8) and the k-order deformation (10), then u(x) is the exact solution of (2).

Proof.

Since the series (17)k=1ui,k(x,C1,C2,,Ck) converges, it can be written as (18)Si(x)=k=1ui,k(x,C1,C2,,Ck), and it holds that (19)limkui,k(x,C1,C2,,Ck)=0. The left hand-side of (10) satisfies (20)ui,1(x,C1)+k=2nui,k(x,Ck)-k=2nui,k-1(x,Ck-1)=ui,2(x,C2)-ui,1(x,C1)++ui,n(x,Cn)-ui,n-1(x,Cn-1)=ui,n(x,Cn). According to (18) we have (21)ui,1(x,C1)+k=2nui,k(x,Ck)-k=2nui,k-1(x,Ck-1)=limnui,n(x,Cn)=0. Using the linear operator Li, (22)Li(ui,1(x,C1))+k=2Li(ui,k(x,Ck))-k=2Li(ui,k-1(x,Ck-1))=Li(ui,1(x,C1))+Lik=2ui,k(x,Ck)-Lik-1ui,k-1(x,Ck-1)=0 which satisfies (23)Li(ui,1(x,C1))+Lik=2ui,k(x,Ck)-Lik-1ui,k-1(x,Ck-1)=k=2[m=1k-1CkNi,0(ui,0(x))+m=1k-1Cm[Li(ui,k-m(x,Ck-m))+Ni,k-1-m(ui,k-1)(ui,k-m(x,Ck-m))]m=1k-1]+gi(x)=0. Also the right hand side can be written as (24)k=1[m=1kCm-k×[Li(ui,m-1(x,Cm-1))+Ni,m-1(ui,k-1)]m=1k]+gi(x)=0. Now, if the Cm,m=1,2,, is properly chosen, then (24) leads to (25)Li(ui(x))+Ni(ui(x),ui(ξj(x)))+gi(x)=0, which is the exact solution.

4. Applications

In this section, we will present a few examples with a known analytic solution in order to demonstrate the effectiveness and high precision of this algorithm.

Example 1.

Consider the following linear delay differential equation : (26)u(x)=12ex/2u(x2)+12u(x),0x1,u(0)=1, with the exact solution (27)u(x)=ex. Applying the procedure which is described in Section 2, the linear and nonlinear operators are (28)L[v(x,p)]=dv(x,p)dx,N[v(x,p)]=dv(x,p)dx-12ψ(x;p)v(x2,p)-12v(x,p),v(0,p)=1, where ψ(x;p) is the expansion Taylor series of ex/2 with respect to p, which can be written as (29)ψ(x;p)=k=01k!(x2)kpk.

Now, apply (4) to p=0 to give the zeroth-order problem as (30)u0(x)=0,u0(0)=1. The solution of the zeroth-order deformation is (31)u0(x)=0. The first-order deformation which is obtained from (8) is given as (32)u1(x,C1)=-14(4+x)C1,u0(0)=0, and has the solution (33)u1(x,C1)=-C1x-18C1x2. The second-order deformation is given by (9): (34)u2(x,C1,C2)=u1(x,C1)-C2-14xC2-116x2C2-12C1u1(x2,C1)-14xC1u1(x2,C1)-116x2C1u1(x2,C1)-12C1u1(x,C1)+C1u1(x,C1), with initial condition (35)u2(0)=0. The solution of (34) is given by (36)u2(x,C1,C2)=-C1x-18C1x2-C12x+14C12x2+13192C12x3+5512C12x4+12560C12x5-C2x-18C2x2-148C2x3. According to (10), the third-order deformation is defined as (37)u3(x,C1,C2,C3)=u2(x,C1,C2)-C3-14C3x-116C3x2-196C3x3-12C2u1(x2,C1)-14C2xu1(x2,C1)-116C2x2u1(x2,C1)-196C2x3u1(x2,C1)-12C2u1(x,C1)-12C1u2(x2,C1,C2)-14C1xu2(x2,C1,C2)-116C1x2u2(x2,C1,C2)-196C1x3u2(x2,C1,C2)-12C1u2(x,C1,C2)+C2u1(x,C1)+C1u2(x,C1,C2), with initial condition (38)u3(0)=0, and has the solution (39)u3(x,C1,C2,C3)=-C1x-18C1x2-2C12x+12C12x2+1396C12x3+5256C12x4+73840C12x5+118432C12x6-C13x+58C13x2+11192C13x3+174096C13x4-199245760C13x5-3771474560C13x6-1095898240C13x7-77864320C13x8-170778880C13x9-C2x-18C2x2-148C2x3-2C1C2x+12C1C2x2+1196C1C2x3+231024C1C2x4-237680C1C2x5+536864C1C2x6+1258048C1C2x7-C3x-18C3x2-148C3x3-1384C3x4. By using (31), (33), (36), and (39), the third-order approximate solution by OHAM for p=1 is (40)u~(x,C1,C2,C3)=u0(x)+u1(x,C1)+u2(x,C1,C2)+u3(x,C1,C2,C3). By using the proposed method of Section 2 on [0,1], we use the residual error: (41)R=u~(x,C1,C2,C3)-12ex/2u~(x2,C1,C2,C3)-12u~(x,C1,C2,C3). The Less Square error can be formed as (42)J(C1,C2,C3)=01R2dx,J(C1,C2,C3)C1=J(C1,C2,C3)C2=J(C1,C2,C3)C3=0. Thus, the following optimal values of Ci’s are obtained: (43)C1=-1.1862449850,C2=0.0255300261,C3=-0.0070171914. In this case, our approximate solution is (44)u~(x,C1,C2,C3)=1+1.00397x+0.479341x2+0.194878x3+0.0349963x4+0.00455719x5+0.00050723x6+0.0000309654x7+1.4858×10-6x8+2.35841×10-8x9. Equations (44) and (41) are plotted in Figures 1(a) and 1(b), respectively. Figure 1(a) shows a comparison between the approximate solution which is obtained by using OHAM and exact solution (27). The residual error is plotted in Figure 1(b). We noted that the absolute maximum error for solving this example via HAM is 0.04 while the absolute maximum error via OHAM is 0.2×10-3, which leads to conclude that OHAM is more accurate than HAM.

(a) Comparison between the three-term OHAM approximate solution and the exact solution for Example 1 and (b) residual error R(x) given by (41) using the three-term OHAM approximate solution.

Example 2.

Consider the linear delay differential equation of third order  (45)u′′′(x)=-u(x)-u(x-0.3)+e-x+0.3,0x1,u(0)=-1,u′′(0)=1,u(x)=e-x,x0, with exact solution (46)u(x)=e-x. According to the method which was described in the above section, we start with (47)L[v(x,p)]=d3v(x,p)dx3,N[v(x,p)]=d3v(x,p)dx3+v(x,p)+v(x-0.3,p)-e-x+0.3. By applying OHAM, we have the following zero-, first-, second-, and the third-order approximate solutions: (48)u0(x)=1-x+0.5x2,(49)u1(x,C1)=0.174167C1x3-0.0541667C1x4+0.0166667C1x5,(50)u2(x,C1,C2)=0.174167C1x3-0.0541667C1x4+0.0166667C1x5+0.173303C12x3-0.0519354C12x4+0.0134917C12x5+0.00356944C12x6-0.000634921C12x7+0.0000992063C12x8+0.166667C2x3-0.0416667C2x4+0.00833333C2x5,(51)u3(x,C1,C2,C3)=0.174167C1x3-0.0541667C1x4+0.0166667C1x5+0.346606C12x3-0.103871C12x4+0.0269833C12x5+0.00713889C12x6-0.00126984C12x7+0.000198413C12x8+0.172448C13x3-0.0497316C13x4+0.0103718C13x5+0.0070607C13x6-0.00119984C13x7+0.000156374C13x8+0.000017306C13x9-2.09436×10-6C13x10+2.00417×10-7C13x11+0.166667C2x3-0.0416667C2x4+0.00833333C2x5+0.33916C1C2x3-0.0915255C1C2x4+0.0189125C1C2x5+0.00682639C1C2x6-0.00109127C1C2x7+0.00014881C1C2x8+0.165917C3x3-0.0397917C3x4+0.00583333C3x5+0.00138889C3x6. By adding (48)–(50) and (51), we obtain (52)u~(x,C1,C2,C3)=u0(x)+u1(x,C1)+u2(x,C1,C2)+u3(x,C1,C2,C3). Following the procedure described in Section 2 regarding the domain between a=0 and b=1, we use the residual (53)R=u~′′′(x,C1,C2,C3)+u~(x,C1,C2,C3)+u~(x-0.3,C1,C2,C3)-e-x+0.3. The following optimal values of Ci’s are obtained: (54)C1=0.2237074205,C2=-0.5559972626,C3=-0.4753337145. By substituting values in (52), we have (55)u~(x,C1,C2,C3)=1-x+0.5x2-0.161547x3+0.0319249x4-0.00106469x5-0.000894311x6+0.0000269763x7-1.86401×10-6x8+1.93748×10-7x9-2.34472×10-8x10+2.24376×10-9x11. The comparison between the approximate solution and the exact solution is shown in Figures 2(a) and 2(b). We observe that the results agree very well with the exact solution.

(a) Comparison between the three-term OHAM approximate solution and the exact solution for Example 2 and (b) residual error R(x) given by (50) using three-term OHAM approximate solution.

Example 3.

Consider the first order of nonlinear delay differential equation  (56)u(x)=-2u2(x2)+1,0x1,u(0)=0, which has the exact solution (57)u(x)=sin(x). By applying the same method as in Examples 1 and 2, we have the following: (58)L[v(x,p)]=dv(x,p)dx,N[v(x,p)]=dv(x,p)dx+2v2(x2,p)-1. According to OHAM, we have the following zero-, first-, second- and the third-order approximate solutions: (59)u0(x)=x,u1(x,C1)=16x3,u2(x,C1,C2)=16C1x3+16C12x3+1120C12x5+16C2x3,u3(x,C1,C2,C3)=16C1x3+13C12x3+160C12x5+16C13x3+160C13x5+15040C13x7+16C2x3+13C1C2x3+160C1C2x5+16C3x3.

From (59), the third-order approximate solution by OHAM is given as (60)u~(x,C1,C2,C3)=u0(x)+u1(x,C1)+u2(x,C1,C2)+u3(x,C1,C2,C3). By using (60) in (14) and applying the method as discussed in (15) and (16), we obtain the following values of Ci’s: (61)C1=-0.9892887781,C2=0.0001159690,C3=3.3939542758. The approximate solution now becomes (62)u~(x,C1,C2,C3)=x-0.166665x3+0.00832857x5-0.000192105x7. From Figures 3(a) and 3(b), we observe that the results agree very well with the exact solution; as we increase the order of the problem the accuracy increases and the residual error will decrease as shown in Figure 3(b). We observed that the absolute maximum error for solving this example via HAM is 1.2×10-6 while the absolute maximum error by using OHAM is 4×10-8, which revealed that the proposed method is more accurate than HAM.

(a) Comparison between three-term OHAM approximate solution and the exact solution for Example 3 and (b) the residual error for Example 3 using two-term and three-term OHAM approximate solutions.

Example 4.

Consider the third-order nonlinear delay differential equation  (63)u′′′(x)=2u2(x2)-1,0x1,u(0)=0,u(0)=1,u′′(0)=0. The exact solution of the above problem is given as (64)u(x)=sin(x). By applying the present method, the linear and nonlinear operators are defined as (65)L[v(x,p)]=d3v(x,p)dx3,N[v(x,p)]=d3v(x,p)dx3-2v2(x2,p)+1. According to OHAM, we have the following zero-, first-, second- and third-order approximate solutions: (66)u0(x)=x-x36,u1(x,C1)=-1120C1x5+15040C1x7-1580608C1x9,u2(x,C1,C2)=-1120C1x5+15040C1x7-1580608C1x9-1120C12x5+15040C12x7-11451520C12x9-139916800C12x11+1011275293859840C12x13-19738607656960C12x15-1120C2x5+15040C2x7-1580608C2x9,u3(x,C1,C2,C3)=-1120C1x5+15040C1x7-1580608C1x9-160C12x5+12520C12x7-1725760C12x9-119958400C12x11+101637646929920C12x13-14869303829480C12x15-1120C33x5+15040C13x7+12903040C13x9-119958400C13x11  +4916376469299200C13x13+12232678117105664000C13x15-341827149186120241276518400C13x17+15040C3x7-1580608C3x9-26391675269781634049638400C13x21-1120C2x5+15040C2x7-1580608C2x9-160C1C2x5+12520C1C2x7-1725760C1C2x9-119958400C1C2x11+101637646929920C1C2x13-1120C3x5-14869303828480C1C2x15+3803310204330624503313858560C13x19. From (66), the third-order approximate solution by OHAM is given as (67)u~(x,C1,C2,C3)=u0(x)+u1(x,C1)+u2(x,C1,C2)+u3(x,C1,C2,C3). By using (67) in (14) and applying the method as discussed in (15) and (16), we obtain the following values of Ci’s: (68)C1=2.7354549148,C2=-1.3589185618,C3=-1.0003665216. The approximate solution now becomes (69)u~(x,C1,C2,C3)=x-16x3+0.00833639x5-0.000198485x7+1.72296×10-6x9+1.80627×10-29x11-5.71017×10-32x13+7.40356×10-35x15-4.68993×-50x17+7.62895×1053x19-5.87208×10-56x21. Numerical results of the solution are displayed in Figures 4(a) and 4(b).

(a) Comparison between three-term OHAM approximate solution and the exact solution for Example 4 and (b) the residual error for Example 4 using three-term OHAM approximate solution.

Example 5.

Consider the system of delay differential equation  (70)u1(x)=u1(x-1),u2(x)=u1(x-1)+u2(x-0.2),u3(x)=u2(x-1), with initial conditions (71)u1(0)=1,u2(0)=1,u3(0)=1. Following the same procedure, we have (72)Li[vi(x,p)]=dvi(x,p)dx,i=1,2,3,N1[v1(x,p)]=dv1(x,p)dx-v1(x-1,p),N2[v2(x,p)]=dv2(x,p)dx-v1(x-1,p)-v2(x-0.2,p),N3[v3(x,p)]=dv3(x,p)dx-v2(x-1,p). According to OHAM formulation, we have the following:

zeroth-order solution: (73)u1,0(x)=1,u2,0(x)=1,u3,0(x)=1,

first-order solution: (74)u1,1(x)=-C1x,u2,1(x)=-2K1x,u3,1(x)=-A1x,

second-order solution: (75)u1,2(x)=-C1x-2C12x+12C12x2-C2x,u2,2(x)=-2K1x-3.4K12x+32K12x2-2K2x,u3,2(x)=-A1x-3A12x+A12x2-A2x.

Making use of (73)–(75) and extending the solutions up to a fifth order, the approximate solutions by OHAM for p=1 are (76)u~1(x,C1,C2,C3,C4,C5)=u1,0(x)+u1,1(x,C1)+u1,2(x,C1,C2)+u1,3(x,C1,C2,C3)+u1,4(x,C1,C2,C3,C4)+u1,5(x,C1,C2,C3,C4,C5),u~2(x,K1,K2,K3,K4,K5)=u2,0(x)+u2,1(x,K1)+u2,2(x,K1,K2)+u2,3(x,K1,K2,K3)+u2,4(x,K1,K2,K3,K4)+u2,5(x,K1,K2,K3,K4,K5),u~3(x,A1,A2,A3,A4,A5)=u2,0(x)+u2,1(x,A1)+u2,2(x,A1,A2)+u2,3(x,A1,A2,A3)+u2,4(x,A1,A2,A3,A4)+u2,5(x,A1,A2,A3,A4,A5). By using the proposed procedure which is described in Section 2 on [0,1], we use the residual error (77)R1=u1~(x,C1,C2,C3,C4,C5)-u1(x-1,C1,C2,C3,C4,C5),R2=u2~(x,K1,K2,K3,K4,K5)-u1(x-1,C1,C2,C3,C4,C5)-u2(x-0.2,K1,K2,K3,K4,K5),R3=u3~(x,A1,A2,A3,A4,A5)-u2(x-1,K1,K2,K3,K4,K5). The following values of Ci’s, Ki’s and Ai’s are obtained: (78)C1=-0.58668069,C2=-0.08423352,C3=0.00786513,C4=0.00069702,C5=-0.00034937,K1=-0.75856373,K2=-0.14305522,K3=0.07154350,K4=-0.02323472,K5=-0.00280683,A1=-0.67739965,A2=-0.08965635,A3=0.03646196,A4=-0.01330410,A5=0.00453433. By using the above values, the approximate solutions are (79)u~1(x,C1,C2,C3,C4,C5)=1+0.567178x+0.160857x2+0.0302719x3+0.00434813x4+0.000579199x5,u~2(x,K1,K2,K3,K4,K5)=1+1.3299x+0.714785x2+0.205164x3+0.0536786x4+0.0125583x5,u~2(x,A1,A2,A3,A4,A5)=1+0.182376x+0.230372x2+0.0785302x3+0.0228859x4+0.00594311x5. From Figure 5, we can observe the accuracy of the solution obtained by the five-term approximate solution using OHAM which is quite good.

The residual errors for Example 5 using the five-term OHAM approximate solution.

5. Conclusions

In this work, OHAM is employed for the first time to propose a new analytic approximate solution of delay differential equations (DDEs). This method has been tested in various examples of linear and nonlinear and system of initial value problems of DDEs and was seen to yield satisfactory results. The OHAM provides us with a simple way to optimally control and adjust the convergence solution series and it gives a good approximation in few terms which is converged to the exact solution and proved the efficiency and reliability of the method. This fact is obvious from the use of the auxiliary function H(p). In OHAM, it is important to solve a set of nonlinear algebraic equations with m unknown convergence control parameters, C1,C2,,Cm, and this makes it time consuming, especially for large m.

Patel N. K. Das P. C. Parbhu S. S. Optimal control of systems described by delay differential equations International Journal of Control 1982 36 2 303 311 10.1080/00207178208932894 Glass L. Mackey M. C. Pathological conditions resulting from instabilities in physiological control systems Annals of the New York Academy of Sciences 1979 316 214 235 2-s2.0-0018373214 10.1111/j.1749-6632.1979.tb29471.x ZBL0427.92004 Busenberg S. Tang B. Mathematical models of the early embryonic cell cycle: the role of MPF activation and cyclin degradation Journal of Mathematical Biology 1994 32 6 573 596 2-s2.0-0028311386 10.1007/BF00573462 ZBL0823.92004 Lv C. Yuan Z. Stability analysis of delay differential equation models of HIV-1 therapy for fighting a virus with another virus Journal of Mathematical Analysis and Applications 2009 352 2 672 683 2-s2.0-57749096208 10.1016/j.jmaa.2008.11.026 ZBL1160.37437 Evans D. J. Raslan K. R. The Adomian decomposition method for solving delay differential equation International Journal of Computer Mathematics 2005 82 1 49 54 2-s2.0-27944439311 10.1080/00207160412331286815 ZBL1069.65074 Saeed R. K. Rahman B. M. Adomian decomposition method for solving system of delay differential equation Australian Journal of Basic and Applied Sciences 2010 4 8 3613 3621 2-s2.0-78650279514 Shakeri F. Dehghan M. Solution of delay differential equations via a homotopy perturbation method Mathematical and Computer Modelling 2008 48 3-4 486 498 2-s2.0-44949134839 10.1016/j.mcm.2007.09.016 ZBL1145.34353 Koçak H. Yildirim A. Series solution for a delay differential equation arising in electrodynamics Communications in Numerical Methods in Engineering 2009 25 11 1084 1096 2-s2.0-77649175072 10.1002/cnm.1288 ZBL1177.78059 Vanani S. K. Aminataei A. Multiquadric approximation scheme on the numerical solution of delay differential systems of neutral type Mathematical and Computer Modelling 2009 49 1-2 234 241 2-s2.0-55549132926 10.1016/j.mcm.2008.03.016 ZBL1165.65366 Vanani S. K. Aminataei A. On the numerical solution of neutral delay differential equations using multiquadric approximation scheme Bulletin of the Korean Mathematical Society 2008 45 4 663 670 2-s2.0-57349141862 10.4134/BKMS.2008.45.4.663 ZBL1168.65039 Saadatmandi A. Dehghan M. Variational iteration method for solving a generalized pantograph equation Computers and Mathematics with Applications 2009 58 11-12 2190 2196 2-s2.0-70350575223 10.1016/j.camwa.2009.03.017 ZBL1189.65172 Rangkuti Y. M. Noorani M. S. M. The exact solution of delay differential equations using coupling variational iteration with Taylor series and small term Bulletin of Mathemaatics 2012 4 1 1 15 El-Safty A. Salim M. S. El-Khatib M. A. Convergence of the spline function for delay dynamic system International Journal of Computer Mathematics 2003 80 4 509 518 2-s2.0-27944451151 10.1080/0020716021000014204 ZBL1022.65075 Alomari A. K. Noorani M. S. M. Nazar R. Solution of delay differential equation by means of homotopy analysis method Acta Applicandae Mathematicae 2009 108 2 395 412 2-s2.0-70349884138 10.1007/s10440-008-9318-z ZBL1187.34081 Sedaghat S. Ordokhani Y. Dehghan M. Numerical solution of the delay differential equations of pantograph type via Chebyshev polynomials Communications in Nonlinear Science and Numerical Simulation 2012 17 12 4815 4830 10.1016/j.cnsns.2012.05.009 Nandakumar K. Wiercigroch M. Galerkin projections for state-dependent delay differential equations with applications to drilling Applied Mathematical Modelling 2013 37 4 1705 1722 10.1016/j.apm.2012.04.038 Hafshejani M. S. Vanani S. K. Hafshejani J. S. Numerical solution of delay differential equations using Legender wavelet method World Applied Sciences Journal 2011 13 27 33 Karakoc F. Bereketolu H. Solutions of delay differential equations by using differential transform method International Journal of Computer Mathematics 2009 86 5 914 923 2-s2.0-67651152889 10.1080/00207160701750575 ZBL1167.65038 Li D. Liu M. Z. Runge-Kutta methods for the multi-pantograph delay equation Applied Mathematics and Computation 2005 163 1 383 395 2-s2.0-12444283754 10.1016/j.amc.2004.02.013 ZBL1070.65060 Marinca V. Herişanu N. Bota C. Marinca B. An optimal homotopy asymptotic method applied to the steady flow of a fourth-grade fluid past a porous plate Applied Mathematics Letters 2009 22 2 245 251 2-s2.0-56949103791 10.1016/j.aml.2008.03.019 ZBL1163.76318 Marinca V. Herişanu N. Nemeş I. Optimal homotopy asymptotic method with application to thin film flow Central European Journal of Physics 2008 6 3 648 653 2-s2.0-48349119574 10.2478/s11534-008-0061-x Marinca V. Herişanu N. Application of optimal homotopy asymptotic method for solving nonlinear equations arising in heat transfer International Communications in Heat and Mass Transfer 2008 35 6 710 715 2-s2.0-43449134670 10.1016/j.icheatmasstransfer.2008.02.010 Herisanu N. Marinca V. Dordea T. Madescu G. A new analytical approach to nonlinear vibration of an electric machine Proceedings of Romanian Academy A 2008 9 3 Herişanu N. Marinca V. Explicit analytical approximation to large-amplitude non-linear oscillations of a uniform cantilever beam carrying an intermediate lumped mass and rotary inertia Meccanica 2010 45 6 847 855 2-s2.0-78650716870 10.1007/s11012-010-9293-0 ZBL1258.74102 Islam S. Ali Shah R. Ali I. Allah N. M. Optimal homotopy asymptotic solutions of couette and poiseuille flows of a third grade fluid with heat transfer analysis International Journal of Nonlinear Sciences and Numerical Simulation 2010 11 6 389 400 2-s2.0-78649501970 Idrees M. Islam S. Haq S. Islam S. Application of the optimal homotopy asymptotic method to squeezing flow Computers and Mathematics with Applications 2010 59 12 3858 3866 2-s2.0-77953321597 10.1016/j.camwa.2010.04.023 ZBL1205.76189