International Refereed Journal of Engineering and Science (IRJES)

ISSN (Online) 2319-183X, (Print) 2319-1821

Volume 2, Issue 2(February 2013), PP.45-49
www.irjes.com
www.irjes.com 45 | Page
Variational Formulation For Linear and nonlinear Problems

Dr.Sami. H.Altoum
(Ph.D Mathematics)
Umm Al-qura University KSA-University College of Al- Qunfudah


Abstract: This paper introduced the method of variational formulation for linear and nonlinear problems, and
we show This method is suitable for equations of fluid dynamics since it is applicable to nonlinear evolutionary
types of equations. We show that the variational characterization of a problem depends on the form of the
equation underlying the problem.
Keyword: Gateaux Derivatives, Symmetry, Variational Problem.

I. Introduction:
The problem of giving variational formulation for all linear or nonlinear problems has so far eluded
solution. As a matter of fact,variational principle do not exist for many heat and mass transfer problems of
intrest. We examine the question of existence of variational principle in a more systematic way. We first
introduce the notion of Gateaux differentials and Gateaux derivatives in order to be able to give a general
treatment of variational formulation for nonlinear differential equation.

1-Gateaux Derivatives:
Consider a vector field v,we derive v from a potential ,if v = 0 ,then v can be represented as the
gradient of a potential,v = .It is clear from this elementary concept in vector calculus that if we regard the
Euler equation in variational principle as the gradient of a functional ,analogous to a potential ,then we should
not expect every differential equation to be derivable from a potential .We express in the general form

( )
v
u N | = (1)

Where N denotes an operator which may by nonlinear. The set of elements u that satisfies the given initial or
boundary conditions and the given functional class is called the domain of operator and is denoted by ( ) N D
which can be considered as a subset of a vector spaceU .The set of elements ( ) u N v constitutes the range
of N denoted by ( ) N R .Suppose limit below exist

( )
( ) ( )
( )
0 0
lim
=
+
c
c
=
+
= '
c c
c|
c c
c|
| u N
u N u N
N
u
(2)

Equation (2) is defined by the topology of the space V .The
u
N'
is called the Gateaux differential of the
operator N in the direction | and
u
N'
is called Gateaux derivative of the operator N .Now we define a bilinear
functional ) ( u v, ,which is map 9 U V B: .The expression ) ( u v, satisfies the following requirement :

i. It must be real-valued even if U andV are vector space over the complex number field.
ii. It must be bilinear over the real number field.
iii. It must be non-degenerate this implies that
If 0 ,
0
= ) ( u v ,for every V v e then
u
u | =
0
;
If 0 ,
0
= ) ( u v ,for every U u e ,then
v
v | =
0

Variational Formulation For Linear and nonlinear Problems
www.irjes.com 46 | Page
The real number ) ( = v u S , is called the scalar product of the element V v e and U u e . TheV -
space is called the dual space or the conjugate of the V-space or
-
=U V .In any physical problem we
take the bilinear functions such as

= ) ( u v,
}
V(X)

U(X)
O d (3)
Where O is a subset of
N
R and X is a point of O with X =

( )
n
x x x ,..., ,
2 1

Here V.U is the scalar- product.
2-Symmetry condition:
The symmetry condition defined by the following:

} }
+ ' = ' + dV N dV N
u u
| | (4)
Where v u, and X e + the operator
u
N'
is symmetric. The fact that (4) is the condition for the existence of
function having the operator ( ) u N .
Theorem:
Suppose that the following conditions are satisfied:
i. N is an operator from a normal space E into its conjugate space
-
E .
ii. N has a linear Gateaux differential ( ) h x DN , at any every point of all balls
r x x S <
0
:
.

iii.
The bilinear functional
( ) ( )
2 1
, , h h x DN be symmetric for every x in S

( ) ( ) =
2 1
, , h h x DN ( ) ( )
1 2
, , h h x DN (5)
for every
1
h and
2
h in E and every
x
in
. D
equation (5) is just the symmetry condition(4) if F is the
functional arising in a variational principle ,then we deduce ( ) 0 = u N is the corresponding Euler equation,
consequently ,the equation as whether a variational principle exist for given operator

N depends on whether
the operator has symmetric Gateaux differential expressed by the condition (4),to illustrate this notion we give
the following example:

Example:
Let ( ) 0 , , =
jk j
u u u f
Now from the formula (2) we deduce
( )
0
, ,
=
(

+ + +
c
c
= '
c
c| c| c|
c
jk jk j j u
u u u f N

=
jk
jk
j
j
u
f
u
f
u
f
| | | +
c
c
+
c
c
+
c
c
(6)
To test the symmetry requirement ,we now integrate the following by parts

} }
(
(

c
c
+
c
c
+
c
c
+ = ' + dV
u
f
u
f
u
f
dV N
jk
jk
j
j
u
| | | |

}
+

+ V V
c
c
+ + V
(
(

|
|
.
|

\
|
c
c
V +
c
c
+ +
(
(

|
|
.
|

\
|
c
c
V V +
|
|
.
|

\
|
c
c
V
c
c
dV
u
f
u
f
u
f
u
f
u
f
u
f
j k
jk
j
jk
k
j jk
j k
j
j
2 |
boundary term
Variational Formulation For Linear and nonlinear Problems
www.irjes.com 47 | Page

+ + '
}
dV N
u
|
boundary terms. (7)
Equation (7) defines the Gateaux derivatives u
N'
.which maybe regarded as the ad joint to u
N
.It should be
noted that an ad joint is generally defined for a linear operator but the notion of an ad joint is also usefull for non
linear operator.


3-Variational Formulation
3-1 variational formulation in the restricted sense :
Let a problem ( )
u
u N | = with the operator N (which has be non linear
-
= c U V U N D ) ( ,the
functional F given by:
( ) ) ( = u u N F o o ,
This means that the solutions to the problem are the critical points of F and vice versa ,the operator N :
( ) ( )
-
= c c U V N R U u D be the gradient of the functional and ( ) ' , u N
u
must be symmetric .
3-2 Variational Problem in the Extended Sense:
Let a problem ( )
v
u N | = ,with ( )
-
= c U V U N D N : determine a functional F ,if any ,whose
critical points are solutions to the problem and vice versa. This means that for a given operator N ,there exist an
operator N such that
( ) ( )) ( = u u N F o o ,
And the problems ( )
v
u N | = and ( )
v
u N | = have the same solutions.
Theorem:
Consider the system
( )
v
u N | = (8)
Where N is a nonlinear operator : ( )
-
c c U N R U N D ) ( such that :
i. The solution of the problem exist.
ii. It is a unique.
iii. ) (N D is simply connected.
iv. The Gateaux derivatives ( ) ' , u N
u
exist.
v. ) (
u
N D '
is dense in U
.

vi. ( ) u N
u
-
' is invertible for every ( ) N D u e
.

For every operator K that satisfies the condition (vii)
vii. ( ) ( ) N R K D .
viii. ( ) ( )
-
' c
u
N D K R
ix. It is linear .
x. It is invertible.
xi. It is symmetric the operator N defined by

( ) ( ) ( ) u KN u N u N
u
,
-
' = (9)
It is clear that the solution of (8) is the critical point of the functional
| | ( ) ( )) ( = u KN u N u F ,
2
1
1
(10)
Whose gradient is the N .The functional vanishes when the solution is attained further. If K is positive definite
,then | | u F
1
is the minimum at the critical point. Now for every ( ) K D v e and
v
v | = since | | 0
0 1
= u F , it
follows that
1
F is minimum at
0
u .Note that the linear operator
( ) ( ) Kv u N v u R
u
, ,
-
' = (11)
Variational Formulation For Linear and nonlinear Problems
www.irjes.com 48 | Page
Transforms a given operator N into a potential operator ( ) ( ) Nu u R u N , = ,and hence can be regarded as an
integrating operator.
Example:
Consider the first order nonlinear differential equation with the initial condition given by:
( ) ( ) ( ) t u t f t u , =
, ( ) a u = 0 , ( ) T C u , 0
1
e (12)
Where f is a prescribed function. We obtain a variational problem.

Solution:
This is a Cauchy problem so that its existence and uniqueness are guaranteed under the usual hypothesis. Here
we have ,on comparing with (8) the relation
( )
( )
( ) ( ) ( ) ( )
)
`

e = = T C u a u t u t f
dt
t du
u N , 0 ; 0 ; ,
1
(13)
( ) ( ) ( ) ( ) ( )
)
`

e =
c
c
= ' T C t
u
f
t
dt
d
N
u
, 0 ; 0 0 ;
1
| | | | | (14)
( ) ( ) ( ) ( )
)
`

e + = + +
c
c
+ = + '
-
T A T t
u
f
t
dt
d
N
u
, 0 ; 0 ; (15)
The adjoint homogeneous problem is

( ) ( ) 0 = +
c
c
+ t
u
f
t
dt
d
, ( ) 0 = + T (16)
This equation is linear in + with variable coefficients containing ( ) t u .It can be easily shown that it has only
null solution and so the operator
-
'
u
N is invertible. Thus the conditions of above theorem are fulfilled. Let us
introduce the function h as follows
( ) ( ) ( ) ( ) ( ) ( ) t u t f t u t u t u t h , , , =
. (17)
We know that the operator K appearing in the integral operator (11) should be invertible, symmetric with a
range contained in ( )
-
'
u
N D .These integral operator whose kernel is Green's function of such a differential
operator .Take ,for instance, the operator
( ) ( ) ( )
)
`

e = = = T C u T u u
dt
d
L , 0 , 0 , 0 0 ,
2
2
2

(18)
Which is clearly symmetric and positive definite and invertible. Its inverse is


( ) ( ) ( ) | |
}
+ =
T
vd T t H t Kv
0
t t t t (19)
Where ( ) t H is Heaviside unit step function with the Green's function

( ) ( ) ( ) ( ) t t t t + = T t H t t g , (20)

Using (10) ,we now have the variational formulation of the problem (12) with the required functional given by

( ) ( ) ( ) ( ) ( ) ( ) ( ) ( )
} }
=
T T
dt d u u t h t g t u t u t h u F
0 0
1
, , , , ,
2
1
t t t t
(21)
Where h and g are given by (16) and (20) . Further using (11) the integrating operator ( ) v u R , is given by

Variational Formulation For Linear and nonlinear Problems
www.irjes.com 49 | Page
( ) ( ) ( ) ( ) | | ( )
}
+
(

c
c
=
T
d v T t H t
u
f
dt
d
v u R
0
, t t t t t
Which after integrating by parts, reduce to
( ) ( ) ( ) ( ) ( ) | | ( )
} }
+
c
c
=
T T
d v T t H t
u
f
d v v u R
0 0
, t t t t t t t
If we use Green's function for the operator
( ) ( ) ( )
)
`

e = = = T C u T u u
dt
d
L , 0 , 0 , 0 0 ,
2
2
2

Instead of (18) in (21) and perform two integrations by parts, we obtain the functional
( ) ( ) ( ) ( ) ( ) ( )
} } } }
c
c
+
c c
c
=
T T T T
dt d u
g
t u t f dt d u
t
g
t u u F
0 0 0 0
2
1
,
2
1
t t
t
t t
t
( ) ( ) ( ) ( ) ( )
} }
+
T T
o
dt d u f t g t u t f
0
, , ,
2
1
t t t t (22)
Where ( ) t , t g is given by
( ) ( ) ( ) ( )
T
t
T t H t t g t t t t + = , (23)
It is interesting to note that ( ) u F
1
in (22) is free from ( ) t u which is absorbed by the kernel ( ) t , t g .

Conclusion:
1- This method suitable for equations of fluid dynamics since it is applicable to nonlinear
evolutionary types of equations.
2- This method provide us a procedure for obtaining the required functional for the both
homogeneous and nonhomogeneous boundary conditions
3- We show that the variational characterization of a problem depends on the form of the equation
underlying the problem.
References:
1- Abraham Albert Ungar Analytic Hyperbolic Geometry and Albert Einstein's Special Theory Relativity-World Scientific publishing
Co-Pte.Ltd, (2008).
2- Al Fred Grany, Modern Differential Geometry of Curves and Surfaces with Mathematica, CRC Press (1998).
3- Aubin Thierry, Differential Geometry-American Mathematical Society(2001).
4- Aurel Bejancu &Hani Reda Faran-Foliations and Geometric Structures, Springer Adordrecht, the Netherlands(2006).
5- Bluman G.W & Kumei. S , Symmetry and Differential Equations New York :Springer-Verlag (1998).
6- David Bleaker-Gauge Theory and Variational Principle, Addison- Wesley Publishing Company, (1981).
7- Differential Geometry and the calculus of Variations, Report Hermann-New York and London, (1968).
8- Edmund Bertschinger-Introduction to Tenser Calculus for General Relativity, (2002).
9- M. Lee. John-Introduction to Smooth Manifolds-Springer Verlag, (2002).
10- Nail. H. Ibragimov: Elementry Lie Group Analysis and Ordinary Differential Equations. John Wiley Sons New York, (1996).
11- Olever. P.J, Application of Lie Groups to Differential Equations. New York Springer-Verlag, (1993).
12- Polyanin, A, D. and Zaitsev, V, F, Handbook of Nonlinear Partial Differential Equations, Chapman & hall//CRC Boca, Raton,
(2004).
13- Vladimir. G. Invancevic & Tijana. T. Invancevic Applied Differential Geometry-World Scientific publishing Co-Pte. Ltd, (2007).
14- Torsten Asselmeyer Maluga & Carl. H. Brans-Differential Topology and Space-time Models-World Scientific publishing Co-Pte. Ltd,
(2007).
15- Manuel De Leon Analytical Mechanics-Elserviier Science Publisher, (1989).
16- Serge Lang-Fundamental of Differential Geometry-Springer-Verlage, (1999).



Dr. Sami Hajazi Mustafa received his PhD degree in Differential Geometry from Alneelain
University, Sudan in 2007. He was a head of Department of Mathematics in Academy of
Engineering Sciences in Sudan. Now he is Assistant Professor of Mathematics, University
college of Alqunfudha, Umm Al Qura University , Saudi Arabia.