You are on page 1of 3

For the following type of calculus of variations problems: when you need to find some curve that

minimizes an integral F ( x, y ( x ) , y ' ( x ) ) dx with boundary conditions y ( a ) = ya and


a

y ( b ) = yb (and you also dont have any other conditions or constraints), you can use the
following formula.
F d F

=
0
y dx y '
However there are a lot of tricks you can use:
If F is only a function of y ( x ) and y ' ( x ) then you can expand

F d F

=
0 into:
y dx y '

F
F
F

y '
y '' =
0
y y ' y
y ' y '
Multiply everything by y ' ( x ) :
F
F
F
y '
y '2
y ' y '' =
0
y
y ' y
y ' y '
F
d
0
F y'
=
dx
y '
Integrating once:
F y'

F
=
C
y '

x ) ) y 2 1 + y '2
In our case, when F ( x, y ( x ) , y ' (=

1+ y ' y '
2

y 2 1 + y '2
y '

1+ y '
2

y 2 y '2
1 + y '2

)=
C

=
C

Keep simplifying:
y 2 (1 + y '2 )
1 + y '2

y 2 y '2
1 + y '2

=
C

1 + y '2 y '2
=C
y2
1 + y '2

y2
1 + y '2

=C

=
y 4 C 2 (1 + y '2 )

y4
1 =y '
C2
Even without integrating this, you can see that y = 0 is a solution to this differential equation
that satisfies the boundary conditions. And isnt there a theorem in DE that says the solution to a
first order differential equation given at least one initial/boundary condition is unique?
Then you can just say that a straight line between (0,0) and (1,0) is the y(x) we are looking for.

For problem number 4 which has the following integrand:

F ( x, y ( x ) , y ' ( x ) ) = y '2 + y 2 + y
Lets apply the same trick.
F y ' Fy ' =
C

(y' + y
2

+ y) y '

(y' + y
2

( y '2 + y 2 + y )
y '

=
C

+ y ) 2 y '2 =
C

y '2 + y 2 + y =
C
y' =

y '=

Lump C

C + y2 + y

1
1
C + y+
4
2

1
into some constant 1/ K 2 . Substitute
=
K ( y + 1/ 2 ) and ' = Ky '
4

'
=
K
='

1 2
+
K2 K2
1+ 2

=
sinh ( x + C1 )
y=

1 sinh ( x + C1 ) 1
=

2
K
2

There is actually no solution to this problem. There is no C1 and K such that it can satisfy the
boundary conditions of y(0) = 0 and y(1) = 0 and the above equation in the box.

You might also like