Alexandria Engineering Journal (2013) 52, 801–805

Alexandria University

Alexandria Engineering Journal

www.elsevier.com/locate/aej
www.sciencedirect.com

ORIGINAL ARTICLE

A new type of shooting method for nonlinear boundary

value problems
Muhammad Ahsan *, Sarah Farrukh

School of Chemical and Materials Engineering, National University of Sciences and Technology, H-12, Islamabad 44000, Pakistan

Received 11 April 2013; revised 16 July 2013; accepted 18 July 2013

Available online 12 August 2013

KEYWORDS
Boundary value problem;
New shooting method;
Nonlinear boundary value
problems;
Initial value problems;
Interpolation
Abstract In this article we introduce a new type of iterative method for initial value problems
(IVPs). We enhance this method by using shooting techniques and interpolation for the boundary
value problems. Our method is more accurate and applicable than built in methods used in different
software packages. We solved several examples for initial value problems and linear and non-linear
boundary value problems and compared results to those obtained using MATLAB.
ª 2013 Production and hosting by Elsevier B.V. on behalf of Faculty of Engineering, Alexandria
University.

1. Introduction who developed and analyzed both a simple shooting method

Many techniques for solving boundary value problems were
discussed and presented by many researchers. The common
technique for solving boundary value problems is shooting
method. In shooting method the boundary value problem
(BVP) is reduced to the solution of an initial value problem
(IVP), by assuming initial values that would have been given
if the ordinary differential equation were an initial value prob-
lem. The boundary value calculated is then matched with the
real boundary value. Using trial and error or some scientific
approach, one attempts to get as close to the boundary value
as possible [1,2]. The concept of multiple shooting was first
suggested by Morrison et al. [3] later promoted by Keller [4]
* Corresponding author. Tel.: +92 3336057937.
E-mail address: ahsan@scme.nust.edu.pk (M. Ahsan).
Peer review under responsibility of Faculty of Engineering, Alexandria
University.
(SSM) and a multiple shooting method (MSM). A more latest
version of a multiple shooting method, MUSN, was developed
by Mattheij and Staarink [5]. Lie-group shooting method was
proposed by Liu [6–8].
In this article, we develop an iterative formula for the ini-
tial value problem (IVP) and convert it to boundary value
problem by shooting technique and interpolation. The
mostly used shooting technique in different software depends
on Newton Raphson method, which fails to predict results
when the first derivative of function is zero or undefined
.Thus, technique fails in many problems, while we used
interpolation to approximate the guess, due to this we can
solve several problems using shooting method that could
not be solved. This article is organized as follows. Section 1
is an introduction and Section 2 introduces a new approxi-
mating formula for the IVP. Section 3 describes the shooting
technique and in Section 4 some nonlinear problems are
solved using the proposed shooting method. Section 5 con-
cludes the paper.

Production and hosting by Elsevier

1110-0168 ª 2013 Production and hosting by Elsevier B.V. on behalf of Faculty of Engineering, Alexandria University.
http://dx.doi.org/10.1016/j.aej.2013.07.001
802 M. Ahsan, S. Farrukh

2. New approximating formula for initial value problem Comparing (2) and (9)
AþBþC¼1
Let us consider the nonlinear initial value problem
Let
y0 ¼ fðx; yÞ; yðxo Þ ¼ yo
A ¼ B ¼ C ¼ 1=3
We consider the Taylor series expansion of yðx þ hÞ
1 1 BP þ CR ¼ 1=2
yðx þ hÞ ¼ yðxÞ þ hy0 ðxÞ þ h2 y00 ðxÞ þ h3 y000 ðxÞ
2 3!
þ Oðh4 Þ ð1Þ BQ þ CS ¼ 1=2

where the given nonlinear part of the initial value problem is CR2 ¼ 1=3
y0 ðxÞ ¼ fðx; yÞ
1=3ðP þ RÞ ¼ 1=2 ) P þ R ¼ 3=2 ) P ¼ 1=2
Now, we can further approximate for second and third
derivatives, 1=3ðQ þ SÞ ¼ 1=2 ) Q þ S ¼ 3=2
y00 ðxÞ ¼ fx ðx; yÞ þ fy ðx; yÞy0 ðxÞ
R2 ¼ 1 ) R ¼ 1
00
y ðxÞ ¼ fx ðx; yÞ þ fy ðx; yÞ  fðx; yÞ Solving above equations for the constants and putting val-
000 0 0 2
ues in Eqs. (4)–(8) for solving (3), we have our required for-
y ðxÞ ¼ fxx ðx; yÞ þ fyx ðx; yÞy ðxÞ þ fyy ðx; yÞðy ðxÞÞ mula as

y000 ðxÞ ¼ fxx ðx; yÞ þ fyx ðx; yÞ  fðx; yÞ þ fyy ðx; yÞf2 ðx; yÞ f0 ¼ fðx; yÞ
 
So putting all derivatives approximations in (1) h
f1 ¼ x þ ; y þ hf0
1 2
yðx þ hÞ ¼ yðxÞ þ hfðx; yÞ þ h2 fx ðx; yÞ
2  
1 2 1 h
þ h fy ðx; yÞfðx; yÞ þ h3 fxx ðx; yÞ f2 ¼ x þ h; y þ f1
2 3! 2
1 1
þ h3 fyx ðx; yÞfðx; yÞ þ h3 fyy ðx; yÞf2 ðx; yÞ ð2Þ h
3! 3! y ¼ yðxÞ þ ðf0 þ f1 þ f2 Þ ð10Þ
3
Consider the equation proposed for our iterative formula;
yðx þ hÞ ¼ yðxÞ þ Ahf0 þ Bhf1 þ Chf2 ð3Þ 3. Shooting method
where
f0 ¼ fðx; yÞ ð4Þ Let us consider the two point non-linear boundary value
problem
f1 ¼ fðx þ Ph; y þ Qhf0 Þ ð5Þ y00 ¼ fðx; y; y0 Þ; yðaÞ ¼ a; yðbÞ ¼ b ðAÞ
f2 ¼ fðx þ Rh; y þ Shf1 Þ ð6Þ Our aim is to convert the above boundary value problem into
initial value problem. In this case, (A) equation can be written
f1 ¼ fðx; yÞ þ Phfx ðx; yÞ þ Qhfy ðx; yÞfðx; yÞ ð7Þ as for IVP.
y00 ¼ fðx; y; y0 Þ; yðaÞ ¼ a; y0 ðaÞ ¼ Unknown
S2 h2
f2 ¼ fðx; yÞ þ Rhfx ðx; yÞ þ Shfy ðx; Þfðx; yÞ þ fyy ¼ s ðSayÞ ðAÞ
2
þ SRh2 fx ðx; yÞ þ R2 h2 fxx ðx; yÞ ð8Þ Now, the goal is to find s for which the solution satisfies the
Our aim is to find the values of the constants A, B, C, P, Q, second boundary condition
R and S, putting (4)–(8) in (3)
yðx þ hÞ ¼ yðxÞ þ Ahfðx; yÞ þ Bhfðx; yÞ þ BPh2 fx ðx; yÞ
þ BQh2 fy ðx; yÞfðx; yÞ þ Chfðx; yÞ þ CRh2 fx ðx; yÞ
h3
þ CSh2 fy ðx; yÞfðx; yÞ þ CS2 fyy ðx; yÞ
2
þ CSRh3 fx ðx; yÞfy ðx; yÞ þ CR2 h3 fxx ðx; yÞ
which implies
yðx þ hÞ ¼ yðxÞ þ hfðx; yÞ½A þ B þ C þ h2 fx ðx; yÞ½BP
þ CR þ h2 fy ðx; yÞ½BQ þ CS þ CS2
h3 h3
 fyy ðx; yÞ þ CSRh3 fx fy þ CR2 fxx ð9Þ
2 2 Figure 1 Cross sectional geometry of a pressure vessel.
Shooting method for nonlinear boundary value problems 803

yðbÞ ¼ b: Giving us two first order differential equations as

The method is explained by an example, (see Fig. 1). du
¼ w; uð5Þ ¼ 0:003873100
dr
Example. Take the case of a pressure vessel that is being tested
in the laboratory to check its ability to withstand pressure. For dw w u
¼  þ 2 ; wð5Þ ¼ not known
a thick pressure vessel of inner radius ‘a’ and outer radius ‘b’, dr r r
the differential equation for the radial displacement ‘u’ of a Let us assume
point along the thickness is given by [9]

d2 u 1 du u du uð8Þ  uð5Þ
þ  ¼0 wð5Þ ¼ ð5Þ  ¼ 0:00026538
dr 85
dr2 r dr r2
Set up the initial value problem.
Suppose that the inner radius a = 500 and the outer radius
b = 800 . Two strain gauges that are bonded tangentially at
du
the inner and the outer radius measure the normal tangential ¼ w ¼ f1 ðr; u; wÞ; uð5Þ ¼ 0:003873100
dr
strain in the pressure vessel as
dw w u
et=r¼a ¼ 0:00077462 ¼  þ 2 ¼ f2 ðr; u; wÞ; wð5Þ ¼ 0:00026538
dr r r
et=r¼b ¼ 0:00038462 The results obtained using the New Type Shooting Method are
summarized in Tables 1 and 2 together with a comparisons to
at the maximum needed pressure. Since the radial displace- those obtained using other standard techniques in literature
ment and tangential strain are related simply by [9].
u
et ¼
r 4. Some nonlinear problems are solved using shooting type
Then method
ujr¼a ¼ 0:00077462  5 ¼ 0:003873100
Problem 1. The opening and closing of small high perfor-
ujr¼b ¼ 0:00038462  8 ¼ 0:003077000 mance hydraulic valves used to control guidance jets in a space
Starting with the ordinary differential equation craft are accomplished by the use of torque motor. The IVP
describing the rotation of the motor shaft is [10].
d2 u 1 du u
þ  ¼ 0; uð5Þ ¼ 0:0038731; uð8Þ ¼ 0:0030770
dr2 r dr r2
Let h00 þ 4h0 þ 3h ¼ cos t; hð0Þ ¼ h0 ð0Þ ¼ 0

du in which h stand for angular displacement of the motor

¼w shaft.The results are shown in Fig. 2.
dr
Then Problem 2. Consider the unidirectional thin film flow of an
dw 1 u incompressible fluid of third grade down an inclined plane of
þ w 2 ¼0 inclination a „ 0 in [11]. It is assumed that the ambient air was
dr r r

Table 1 Comparison of new shooting type method (NSTM) with other methods.
r (in) Exact (in) Euler (in) Runge Kutta (in) NSTM
5 3.8731 · 103 3.8731 · 103 3.8731 · 103 3.8731 · 103
5.75 3.5567 · 103 3.5085 · 103 3.5554 · 103 3.6149 · 103
6.5 3.3366 · 103 3.2858 · 103 3.3341 · 103 3.4522 · 103
7.25 3.1829 · 103 3.1518 · 103 3.1792 · 103 3.3549 · 103
8 3.0770 · 103 3.0770 · 103 3.0723 · 103 3.3070 · 103

Table 2 Error in results of new shooting type method (NSTM) and other methods.
r (in) Exact (in) |et| (%) (Euler) |et| (%) Runge Kutta (in) |et| (%) NSTM
3
5 3.8731 · 10 0 0 0
5.75 3.5567 · 103 1.3731 3.5824 · 102 5.82 · 103
6.5 3.3366 · 103 1.5482 7.4037 · 102 1.15 · 102
7.25 3.1829 · 103 9.8967 · 101 1.1612 · 101 1.70 · 102
8 3.0770 · 103 1.9500 · 103 1.5168 · 101 2.30 · 102
804 M. Ahsan, S. Farrukh

0.2

0.15
velocity
0.1 approximate
exact
0.05

y* (t)
-0.05

-0.1

-0.15

-0.2

-0.25
0 0.5 1 1.5 2 2.5 3 3.5 4
Time

Figure 2 Results of Problem 1.

0.7

0.6 beta=0.4
beta=0.2
beta=0.6
0.5

0.4
y* (t)

0.3

0.2

0.1

0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Time

Figure 3 Results of Problem 2.

6 velocity
approximate
second oder
5

4
y* (t)

0
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
Time

Figure 4 Results of Problem 3.

Shooting method for nonlinear boundary value problems 805

stationary, surface tension negligible and thin film of uniform well-known numerical methods. The work would be extended
thickness d. In the absence of pressure gradient, they obtained to develop a code which works as a function in MATLAB like
the governing nonlinear second-order ODE [11]. ode45, ode23 and ode113. Moreover, this method could be fur-
 2 ther extended to deal with non-linear coupled BVP.
d2 t 6ðb2 þ b3 Þ dt d2 t qg
þ  ¼ 0;
dx2 l dx dx2 l
Acknowledgement
with boundary conditions
dt The authors are grateful to Nayyar Mehmood for his guidance
tð0Þ ¼ Uo ; ¼ 0 at x ¼ d
dx and support during the preparation of this paper.
in which v is the fluid velocity, q the density, l dynamic viscos-
ity, b2 and b3 are material moduli of third order fluid, g the References
acceleration due to gravity and d the thickness of the thin
layer. [1] H.G. Bock, K.J. Plitt, A Multiple Shooting Algorithm for Direct
Defining Solution of Optimal Control Problems, 1983.
[2] C. Kirches, H.G. Bock, J.P. Schlöder, S. Sager, Complementary
Ud y condensing for the direct multiple shooting method, in:
U¼ ; y ¼
v d Modeling, Simulation and Optimization of Complex Processes,
Springer, 2012, pp. 195–206.
 2 2
d2 u du d u [3] D.D. Morrison, J.D. Riley, J.F. Zancanaro, Multiple shooting
þ 6b þ k ¼ 0; method for two-point boundary value problems, Communications
dy2 dy dy2
of the ACM 5 (1962) 613–614.
[4] H.B. Keller, Numerical Solution of Two Point Boundary Value
du
uð0Þ ¼ 0; ¼ 0 at y ¼ 1 Problems, SIAM, 1976.
dy [5] R.M. Mattheij, G. Staarink, An efficient algorithm for solving
The obtained results are depicted in Fig. 3. general linear two-point BVP, SIAM Journal on Scientific and
Statistical Computing 5 (1984) 745–763.
[6] C. Liu, The Lie-group shooting method for nonlinear two-point
Problem 3. A Blasius boundary layer in physics and fluid boundary value problems exhibiting multiple solutions,
mechanics describes the steady two dimensional boundary Computer Modeling in Engineering and Sciences 13 (2006) 149.
layer that forms on a semi-infinite plate which is held parallel [7] C. Liu, Efficient shooting methods for the second-order
to constant unidirectional flow U [12]. ordinary differential equations, Computer Modeling in
Engineering and Sciences 15 (2006) 69.
f000 þ ff00 ¼ 0 [8] C. Liu, The Lie-group shooting method for singularly perturbed
with boundary conditions two-point boundary value problems, Computer Modeling in
Engineering and Sciences 15 (2006) 179.
fð0Þ ¼ 0; f0 ð5Þ ¼ 1 [9] A. Kaw, E. Kalu, Numerical Methods with Applications:
Abridged, in, Lulu, Raleigh, NC, USA, 2011.
where f is the dimensionless stream function.
[10] A. Kharab, R.B. Guenther, An Introduction to Numerical
The results are shown in Fig. 4. Methods: A MATLAB Approach, CRC Press, 2011.
[11] M. Sajid, T. Hayat, The application of homotopy analysis
5. Conclusion method to thin film flows of a third order fluid, Chaos, Solitons
& Fractals 38 (2008) 506–515.
A new type of shooting method is proposed in this study. The [12] H. Schlichting, K. Gersten, Boundary-Layer Theory, Springer,
2004.
results show good competitive performance with other