Changing Independent Variable and Normal Form
Changing Independent Variable and Normal Form
which may be solved by any of the methods already discussed. We shall consider Mfferential Equations
these transformations in this unit. We shall also discuss physical situations where
the governing differential equation is of form (3).
Objectives
After reading this unit, you should be able to
e obtain the solution of second order linear differential equation with variable
coefficient by transforming the independent variable;
e obtain the solution of second order linear differential equation with variable
coefficient by transforming the dependent variable;
0 obtain solutions of linear second order equations modelled for certain physical
situations.
We consider the transformations of the variables under which Eqn. (3) reduces to
integrable form. We first consider the transformation of the independent variable.
d 2~
- + P--dy + q y = r , ...(4)
dx dx
where p, q and r are functions of x alone.
- -- - - ( ) = ( )
d 2~
dx2
d
dx
dy d2y
+ dy d2z
dz
-- dx2
a-
d2y +
dz2
Here p,, ql and rl are functions of x but we can immediately express them as
functions of z by means of the relation (5).
It is not easy to solve Eqn. (7) in the general form given above. In order to solve
dy
it, we put some restrictions either on the coefficient of - or on the coefficient
dz
of y. In other words, as z is arbitrary, we choose z so as to sat~sfycertain
, conditions. Accordingly, let us now discuss the following two possibilities:
(i) q, = a Z (a constant)
In this case, Eqn. (7) reduces to
...(9)
Now, q, = a 2 ,
3 adz =&dx
I
...(10)
If with this value of z, p, also becomes constant then Eqn. (9) can be integrated.
For any other form it is not easy to integrate Eqn. (9).
(ii) p, = 0
In this case, Eqn. (7), reduces to
Now, p, = 0,
d 'z dz
3 - + p --- = 0
dx2 dx
dz
a - = e -5 P dx ...(12)
dx
...(13)
If, with this value of z, ql becomes constant or constant divided by z2, then Eqn.
(1 1) can be integrated. For any other form of q l it is not easy to solve Eqn. (1 1).
We will now illustrate these two cases with the help of a few examples.
dZy 2 dy a2
Example 1 : Solve + -- + -- y = 0,x # 0,
dx x d x x4
Solution : Comparing the given equation with Eqn. (4), we get Second Older Linear
Differential Equations
q l q= - constant
- = cl (say), where cl # 0.
/ dz \2
Solving, we get
With the above value of z, and using Eqns. (8), it can be verified that
- y = Acos f i l z + Bsin f i l z
Hence the solution of the given differential equation is
For this value of z, q, becomes a 2 and using Eqn. (ll), the given equation reduces
to
A.E. is
* m = + ia.
:. C.F = Acosaz + Bsinaz.
We will now take up an example which corresponds to the possibility (ii) above
and in which for a given z, q l becomes constant divided by z2.
Example 4 : Solve (x2D2 - 3xD + 3) y = 2x3 - x2
Solution : The given equation can be written as
Intergrating, we get
Now, q, 9
- x3" 3
= - -
3/16 - 3/16
= = --- - - - (which is of the form of
f- dz \2
x6 x8 x8/16 constant divided by z2)
and r, = = --- -
----=---
f- dz \2 x x5 x6 (4zy4 (42)'I4
Its A.E. is
1
and P. I =
(D - 314) (D - 114) 8
t ,.,-
,- ,.,,,A lnz .,. ~ .,-
1
.. y = Az3I4 + BZ"~ + - (lnz) z3l4 + 2 ~ " ~
fi
= +
A ~ X B~~ X + x3 1 1 1 ~+ x2
where Al, B1, A2 are constants.
You may now try to solve the following exercises.
b) (1 + x)' 7
U Y
+ (1 + X) U Y + y = 4 cos In (1 + x), x >
d 2~ dy
C) - + (tam) - + H H
dx dx
(cos2x) y = 0, - -
2
< x < -
2
Second and Higher Order
Ordinary Differential Equations
E3) Solve $
dx"
+ (1 -
\
I2+
\
x j dx
4x2v e-" = 4(x2 + x3) e-lX. x >0
So far we have solved equations of form (3) or (4) by changing the independent
variable. Sometimes an equation of this form can be transformed into an
integrable type by changing the dependent variable. We now consider such
equations in the next section.
Let y = vz,
where v and z are both functions of x.
Then,
and
Note that Eqn. (17) is same as the original Eqn. (4) with right hand side equated
to zero and y replaced by z and hence we may solve it for z. With this choice of
Second Order Linear
the coefficient of v, Eqn. (16) reduces to
Differential Equations
dv
which is a linear equation in - and its integrating factor is
dx
=,
dv A e-
-- = -
dx z2
I Pdx +
z
e-I pdX [ z r e 1 pdx dx]
If, however, we are not able to obtain a solution of Eqn. (17), then it is sometimes
dv
useful to remove the first derivative, (i-e., the coefficient of - ) from the
dx
transformed Eqn. (16). In other words, we try to choose z satisfying the relation,
where
q1 = q - - -1 dp
and
2 dx 4
I
Eqn. (22) is known as the normal form of Eqn. (4) and transforming Eqn. (4) into
form (22) is called removing the first derivative.
Remark : While using this method we advise you to memorise the values of q1 and
rl in terms of p, q and r and directly write down Eqn. (22) for a given equation of
form (4). This will eventually save the labour of making the substitution in the
given equation and reducing it to form (22).
We now take up a few examples to illustrate the above method.
d 5 - x2 -
Example 5 : Solve - dy + xy = xm+'
dx2 dx
Second and Higber Order Solution : Through inspection, we find that y = x is a solution of the
Ordinary Differential Equations homogeneous equation
dv
The above differential equation is linear in - and its integrating factor is
dx
Hence
The second term on the right hand side of Eqn. (25) cannot be integrated in the
present form. However, the last term of Eqn. (25) is integrable if m = 0 or if m is
any positive multiple of 3.
When m = 0, then we have
J xL
Similarly, when m is any positive multiple of 3, say m = 3n,
where n is positive integer, then Second Order .Linear
Differential Equations
J
Integrating n times, we get
and the corresponding form of Eqn. (25) can then be written down.
d 2 ~ 1 dy
Example 6 : Solve 7 - - - +- (- ~ + x ' / ~ + x=) 0
dx x'l2 dx 4x2
Solution : On comparing the given equation with Eqn. (4), we find that
1 dp 1
Also, from Eqn. (23), q, = q - - - - - p2
2 dx 4
Example 7 : Solve
dx
+ (cos2x) y = 0
d 2~ dy
Example 8 : Solve -- 4x - + (4x2 - 1) y = -3 ex2sinx
dx2 dx
2
Solution : Here p = - 4x, q = 4x2-1, r = - 3eXsinx
Using the method of changing dependent variable by using the substitution
y = vz, we have
and 2
-3eX sinx
rl = = - 3sinx
2
ex
Hence equation giving v is
Its A.E is
m 2 + 1 = 0
= ~ m = * i
:. C.F = Acosx + Bsinx
1 1
P.1 = -(- 3sinx) = (-3) -sinx
D2+1 D2+1
3
= -X COSX
2
3
.'. v = C.F + P.1 = Acosx + Bsinx + - xcosx
2
and the complete primitive of the given .equation is
As we have already mentioned in Sec. 9.1, second order equations are most
important for studying several real life situations in physics and biology. In the
b
next section we discuss a few applications of these equations in physical models.
In this section, we discuss some physical models where the governing differential
equations can be approximated to second order linear differential equations.
Generally, vibrations occur whenever a physical system in stable equilibrium is
disturbed, for then it is subjected to forces in order to restore its equilibrium.
We first consider different aspects of mechanical vibrations.