SMA 335; ORDINARY DIFFERENTIAL EQUATION I

CHAPTER 1

DIFFERENTIAL EQUATIONS OF THE FIRST ORDER

1.1 Definitions

Any relation between the variables x, y and the derivatives dy/dx, d 2y/dx2, ... is called an

ordinary differential equation. The term ordinary distinguishes it from a partial differential

equation which involves partial derivatives.

The order of a differential equation is that of the highest derivative occurring in it; the degree of

a differential equation is that to which the derivative of the highest order is raised when the

equation is expressed in a rational integral form.

Throughout this chapter we shall frequently write y' and y" for dy/dx and d2y/dx2 respectively.

1.2 Formation and solution of differential equations

The equation x2+y2=c2••• (i)

where c is an arbitrary constant or parameter, represents a family of concentric circles with

centres at the origin; c is constant for each circle but varies from circle to circle. Differentiating

(i) with respect to x, we have

2x+2y y'=0,

y'= -x/y (ii)

This is the differential equation of the family of circles given by (i). It expresses a property

common to all the members of the family, viz. that the tangent at the point P(x, y) on anyone of

them is perpendicular to the radius which passes through P. The differential equation is of the

first order.

If we start with a relation Such as

y=ax+b/x (iii)

which contains two arbitrary constants a and b, it is necessary to differentiate twice before we

can eliminate the constants.

We have y' =a-b/x2 (iv)

y" =2b/x3 (v)

and eliminating a and b between (iii), (iv) and (v) We obtain the second order differential

equation

x2y"+xy'=y.

The process of finding a relation from which a given differential equation is derived is known as

solving (or integrating) the differential equation. The above examples suggest that the

differential equation obtained from a relation involving n arbitrary constants will be of the nth

order. Conversely, it may be expected that if we can solve a different equation of the nth order.

the most general solution (or integral) will contain n arbitrary constants. Such a solution is

termed the general solution or the complete primitive. There may, however, be other solutions in

addition to the complete primitive .

If particular values are assigned to the n arbitrary constants, a particular solution, called a

particular integral, is obtained. For example, (i) is the general solution of the differential equation

(ii); the circle x2 + y2 = 4 is a particular integral of the differential equation.

Below are given methods of solution for some of the simpler, commonly occurring types of

ordinary differential equation.

1.3. Equations of the first order and first degree

This chapter is mainly devoted to the methods of solution of differential equations of the first

order and first degree, i.e. to equations which may be written in the alternative forms

M(x, y) +N(x, y)y' =0,

M(x, y)dx+N(x, y)dy=0,

where M and N do not involve derivatives of y. The general solution contains one arbitrary

constant which we shall denote by C .

1.4. Variables separable

If a differential equation of the first order and first degree can be writ ten in the form

X dx+ Y dy=0,

where X is a function of x alone, and Y is a function of y alone, the variables are said to be

separable and the solution is obtained by direct integration:

∫X dx+ ∫Y dy=C.
Example 1

Solve the differential equation (1-x2)1(dy/dx)+1+y2=O.

+ = 0,

whence, on integration, tan-1 y + sin-1 x =C. This is the general solution

Example 2

Find the general solution of the differential equation

xy(1 +x2)(dy/dx) - (1 +y2) =0.

The given equation may be written in the form

- = 0,

– =0

whence t log (1+y2)-log x+ log (1+x2)=log C,

 log {(l+x2)(1+y2)}=log Cx,

and so the general solution is {(I+x2)(1+y2)}=Cx.

Note If logarithm functions occur in the integration of a differential equation, the constant of

integration is frequently written in the form log C in order to simplify the form of the general

solution.

Exercises 1(a)

For brevity y' is written for (dy/dx).

1. Obtain a differential equation of the second order by eliminating the constants A and B

from the equation φ=A/r+B.

2. If a and b are arbitrary constants, find the second order differential equation whose

solution is y ==a sec x+b tan x.

Solve the differential equations in Nos. 3-9:

3. y'= x3/y 6. (x2+1)y'=y2+4.

4. y' =ex tan y 7. xy y'=I+x2+y2+x2y2.

5. y'=y/(x2-4). 8. y y' =ex+2y sin x.

9. 2x3 y'=y2 +3xy2, given that y= 1 when x=1.

1.5. The homogeneous equation

The differential equation

Mdx+N dy=O (i)

is said to be homogeneous when M and N are homogeneous functions, of the same degree in x

and y. In this case M/N is a function of y/x and (i) may be written in the form

dy/dx=ƒ(y/x) (ii)

If y/x=v, so that y=vx and dy/dx=v+x dv/dx, (ii) takes the form

v + x(dv/dx) = f(v),

and, separating the variables, we obtain the equation

which may be integrated as before, the original variable being restored by substituting y/x for v

in the result.

Example 3

Solve the equation dy/dx=(x2+2y2)/xy, and find the particular solution for which y=0 when x=l.

If y =vx, the equation becomes

=
 =

Integrating, we obtain

log (1+v2)=log x+log C,

which may be written {l+v2}=Cx,

i.e, {x2+y2} =Cx2.

If y=0 when x=l, C=l. Hence the particular solution is

(x2+y2) =x2.

1.6. Equations reducible to homogeneous form

The equation = (i)

is not homogeneous but may be reduced to homogeneous form when a/a'≠b/b' by writing

Y=ax+by+c and X=a'x+b'y+c'. Then

 =

and = / =( /(

=( /(

i.e =

This equation is homogeneous in X and Y and may be solved as before.

If a/a' =b/b' the substitution z=ax+by will reduce (i) to an equation in which the variables can be

separated.

Example 4

Solve the equations =

(i) Let Y =4x-2y+4 and X =2x+y-2; then the given equation becomes
 =

and the substitution Y =vX gives

i.e. dv =0

Integrating, we have

log (v2+4v-4)+log X=log C,

X√{v2+4v-4}=C,

-or y2 + 4XY – 4X2 =C2.

Restoring the original variables we obtain

4(2x-y+ 2)2+ 8(2x+y-2) (2x-y+2) - 4(2x+y- 2)2 =C2 and simplifying,

 4x2-4xy-y2+8x+4y=C', where C'= C2+4.

(ii) =

Here, since the coefficients of x, y in the numerator and denominator are in the same ratio, we let

z=2x+3y; then the given equation becomes

( -2)= ,

which leads to =

i.e. (2-3/z)dz=7dx.

Integrating. we have

2z-3 log z=7x+C.

i.e, log (2x+3y) =2y-x-l-C'. (C' = -C/3).

Exercises 1 (b)

Solve the differential equations:

1. x(y-3x)y'=2y2-9xy+8x2.
3. xy2 y'=x3+y3.

4. x(y+4x)y' +y(x+4y) =0.

5. (2x+y) y'=6y-4x.

6. (x+2y-3) y'=2x-y+1.

7. (x+y-2) y'=x+y+2.

8. (2x-y)2y'+(x-2y)2=0.

9. (3x+y+3) y'+2(x+3)=0.

10. (2x-4y-8) y'=3x-5y-9.

11. y' =y/x+x sin (y/x).

1.7. Exact equations

The equation =

Example 4 (i), by the standard method may be written in the form

(4x-2y+4)dx- (2x+y-2)dy=0,

(4x+4)dx-2(y dx+x dy) - (y-2)dy=O,

i.e :x { (2X2+ 4x) -2xy- (ty2-2y) } =0.

Integrating, we have, 2x2+4x-2xy-ty2+2y=C.

The equation = may be solved in this way if b = -a'.

The equation is then said to be exact.

An exact equation is of the form

This can be written dƒ(x, y) and so the solution is ƒ(x, y) = constant.

If the equation

M dx+N dy=0 (i)

is exact, there is a function f(x, y) such that

and

= =

Hence, if (i) is exact,

 =

The converse result is true, but is harder to prove; (ii) is the condition for (i) to be exact.

The method of solving exact equations is illustrated in the following example:

Example 5

Solve the equation (3x2+y+l)dx+(3y2+x+l)dy=0.

In the above notation M=3x2+y+l, N=3y2+x+l, and since

= = 1,

the given equation is exact.

Now if f(x, y)=C is its solution,

M =fx=3x2+y+ 1

and N=fy=3y2+x+1

Integrating (i) with respect to x, we have, for each value of y, f =x3 + xy + x + constant.

But the value of this constant may depend on the value of y and so

f=x3+xy+x+φ'(y) (iii

where φ(y) is a function to be determined.

Differentiating (iii) partially with respect to y, we obtain

fy=x+φ'(y)·

But from (ii) fy=3y2+x+ 1

Hence φ'(y)=3y2+1

and so φ(y) =y3+y,

the addition of an arbitrary constant being immaterial.

Substituting this value in (iii), we have

f(x, y) =x3+xy+x+y3+y;

and so the solution of the given equation is

x3+y3+xy+x +y =C.

Exercises 1 (c)

Show that the following equations are exact and integrate them:

1. x y' +y=ex

2. (1/x) v' -v/x2=2x.

3. (ax +hy +g)dx + (hx +by +ƒ)dy =0

4. (1-cos2x)dy+2ysin2xdx=0.
 5. (3x2+ 2y+ 1) dx+ (2x+ 6y2+2) dy =0.

1.8. Integrating factors

If a differential equation M dx-s-N dy=0 becomes exact when· multiplied throughout by a

suitable factor, this factor is known as an integrating factor.

For example, if we calve the equation

3ydx+2xdy=O (i)

by separating the variables, we obtain

x3y2=C,

and, taking the total differential of this result,

3X2y2 dx + 2x3y dy = 0.

Comparison of this equation with (i) shows that x2y is an integrating factor of (i).

Example 6

Show that the equation (3xy-2ay2)dx+ (x2-2axy)dy=0 has an integrating factor which is a

function of x alone. Solve the equation.

Suppose that the integrating factor is f(x); then the equation

(3xy- 2ay2)f(x)dx+ (x2- 2axy)f(x)dy=0 (i)

is exact, and so
 (3xy-2ay2)f(x) t = (x2-2axy)f(x) -t .

i.e. (3x-4ay)f (x) =(x2-2axy)f'(x) +2(x-ay)f(x), f'(x),

and hence = .

This equation is satisfied if f(x) =x.

Substitution of this value in (i) leads to the exact equation

(3x2y-2axy2)dx+ (x3-2ax2y)dy=0 (ii)

If F(x, y) =constant is the solution, (ii) is identical with

Fx dx+Fy dy=0

whence Fx=3x2y-2axy2 (iii)

and Fy=x3-2ax2y (iv)

From (iii) and (iv) F(x, y) may be determined as in Example 5 and the solution of the given

equation shown to be x3y-ax2y2=C.

Example 7

Show that the equation dx+{I+(x+y) tany}dy=0 has an integrating factor of the form (x+y)n,

where n is a constant. Solve the equation.

If (x+y)n is an integrating factor, the equation

is exact. (x+y)n dx+{I+(x+y) tany}(x+y)ndy=0

Hence (x+y)n = {I+(x+y) tany}(x+y)n

n(x+y)n-l=n(x+y)n-1{1+(x+y) tany}+(x+y)n tany

=(x+y)n-l{n+(n+l)(x+y) tany},

n=n+(n+l)(x+y) tan y},

.'. n=-l.

Hence l/(x+y) is an integrating factor 01 the given equation and the equation

is exact.

Comparison with fx dx+ fy dy=0 gives

fx=l/(x+y) (i

and fy=l/(x+y)+tan y (ii

From (i) and (ii), f(x, y) =log (x+y) +log (sec y) and the solution of the given equation may be

written in the form x+y=C cos y.

I .9. The linear equation of the first order

The equation

(i +Py)=Q

where P and Q are functions of x alone, involves y and dy/dx to the degree only and is known as

the first-order linear equation.

To find a method of solution, we consider first the particular cast when Q= 0. The equation is

then

+Py=0 (ii

which may be written

(l/y)dy+Pdx=0

and integrated to give log y+ ∫Pdx=log C

or ye∫Pdx=C.

If we verify this solution by differentiating with respect to x we get

(iii) )=0

) ( ) , when
 = euP

=Pe∫Pdx

( +Py)

Hence the left-hand side of (iii) 1s that of (ii) multiplied by .o=. It follows that e ∫Pdx is an

integrating factor of (ii) and also of (i).

To solve (i) we multiply each side by e∫Pdx and obtain

+Py) = Qe∫dx

( Qe∫dx

i.e. ∫ Qe∫dxdx +C

The student should verify that it is unnecessary to include an arbitrary constant in ∫Pdx when

determining tile integrating factor.

The following results are useful for expressing the integrating factor in its simplest form :

. elog x = x; en log x = (elog x)n = xn; e-n log x = 1/ en log x = 1/ xn.

Example 8
Solve the equation dy/dx+y cot x=2 cos x.

Here P=cot x, ∫ Pdx=log sin x

e ∫Pdx = sin x

Hence, from (23.1), y sin x= ∫sin 2x dx+C

I.e, sin x=C- cos 2x.

Example 9

Solve the equation y+x(x+ 1) (dy/dx) =x(x+ 1)2e-x2

This equation may be written in the form

+ y= (x+1)e-x2

Here P=I/x-l/(x+ 1),

∫ Pdx=log x-log (x+ 1) =log x/(x+ 1)

and so the integrating factor is x/(x+l).

Hence; yx/(x /1) = ∫xe-x2 dx+C from

= C-
i.e. xy =(x+ 1) (C - ).

Exercises 1( d)

Solve the differential equations given in Nos. 1-15 :

1. y' sin x-2y cos x=ex sin3 x,

2. y' cos x-4y sin x=6 cos2 x sin x.

3. x2y'+xy=logx.

4. y' tan x+2y=x cosec x.

5. x(1 +x) y' -y =3x4.

6. (l-x2) y' +xy= (1-x·)3/2 ecos x sin x

7. y' +xy-x3=0, where y=0 when x=0.

8. (x-2)(x-3) y'+2y=(x-l)(x-2),

9. x(1-x2) y+(3x2+1)y=(1+x)3

Find the solution to the equation. which remains finite as x tends to zero.

10. (x-I) y'-2y=(x-l)4 cos2 x.

11. x2 y' +x(3+2x)y=e-2x+e-3x.

12. xy'+(I+x)y=x sin x.

 13. (l+x2) y'-xy=(I+x2)x2

14. (1+3x) y' + (3-9x)y=3.

15. xy'_y=x3 cos x, with y=0 when x=7T.

16. (i) Reduce the differential equation (x+ l)y y' -y 2 =x to a linear form by writing Z=y 2, and

solve it, given that when x=0, y=l.

ii) solve the equation (2x +y) y –x+2y, by writing y=vx. Given that when x=1, y=0.

1.10. Bernoulli's equation .

The equation

+Py = Q

in which P and Q are functions of x alone and n is constant is known as Bernoulli's equation and

is reducible to linear form.

Suppose first that n ≠ I and divide throughout by yn so that the given equation becomes

+P =Q

The form of this equation suggests the substitution

yl-n=v, (l-n)y-n dy/dx=dv/dx,

which reduces the given equation after multiplication by (1-n) to the linear equation

This equation may be solved by the method used before.

If n= I, the given equation may be solved by separating the variables.

Example 10

Solve the equation x dy/dx+y=y2x2 log x. ]

When this equation is divided throughout by x it is seen to be Bernoulli's equation with n=2.

Proceeding as above, we then divide throughout the equation by y2 and put v= y-l so that

dv/dx=-y-2dy/dx.

Then -

The integrating factor is e-∫dx/x =e-1og x= l/x; hence from (i),

(v/x) = -log x,

v/x=C- (x log x-x),

i e. l/xy=C+x(l-log x).
Exercises 1 (e)

Solve the following differential equations:

1. y'+y cot x= y2 sin 2 x

2. y' =y tan x+y3 tan3 x.

3. y' =x/y+y.

4. 2y'sin x-y cos x=y3 sin x cos x,

Also find the particular solution for which y = -1 when x =!-7T.

5. y' + y =xy3

6. y' =2y tan x+y2 tan2 x.

7. 2 y'-y(2x+1)/(x2 +x +1) =xy3/(1-x)

8. (x2-1) y'+xy=exy-2√(x2-1).

CHAPTER 2

LINEAR DIFFERENTIAL EQUATIONS WITH CONSTANT COEFFICIENTS

2.1. The operator D

In this chapter we shall write Dy to denote dy/dx and Dry to denote d ry/dxr so that the symbol D

represents the differential operator d/dx.

With this notation the rules of differentiation give

D(u+v) =Du+Dv,
Dm(Dny) = Dn(Dmy) = Dm+ny

when m and n are positive integers, and

D(ay) =aDy when a is a constant.

Using these results we write, when a and b are constants,

where any function of D is interpreted as operating on the function which follows it.

Now (D-a)(D-b)y= (D-a)(Dy-by)

= D2y-(a+b) dy +aby

={D2- (a+b)D+ab}y

and, similarly, (D-b)(D-a)y={D2-(a+b)D+ab}y.

Hence a second-order operator with constant coefficients may be factorized in the algebraic

sense into two linear factor-;Whose order of arrangement is immaterial. For example

(3+2D-D2)(3 cos x+3 sin x)=3(3-D)(1+D)(cos x+sin x)

=3(3-D)(2 cos x)

=6(3 cos x+sin x).

The same result is obtained from 3(1+D)(3-D)(cos z -l-sin x).

In the same way any function which contains only positive integral powers of D may be

factorized and the factors arranged in any order.

2.2 Applications of the operator D

(a) Operators applied to the Junction ekx where k is constant.

By differentiation, Dekx=kekx,

Dnekx=knekx, when n is a positive integer,

φ(D)ekx –φ(k)ekx

Thus, for example (D3[D2+2)e3x= (27 -9+2)e3x=20e3x.

(b) Operators applied to products of the form e kxV, where k is a constant and V is a Junction of x.

By the product rule for differentiation

D(ekxV) = ekx(D + k) V

and, more generally, when n is a positive integer, Leibniz's theorem gives

Dn(ekxV) =ekx(D+k)nV.

Hence, if φ(D) is any polynomial in D,

φ(D) (ekxV) = ekxφ(D +k)V.

For example, (D -I)(e3x sin x) = e3x(D+ 2) sin x

=e3x(cos x+2 sin x)

and (D2-4){e2x(x2+ I)}

 =e2x(8x+2).

2.3 The linear equation with constant coefficients

In a linear differential equation, y and its derivatives occur separately in the first degree only, and

the linear equation of order n with constant coefficients is of the form

(Dn + a1Dn-l + a2Dn-2 + ., . + an-1D + an)y=ƒ(x) (i)

where the a's are given constants and ƒ(x) is a given function of x.

Throughout this chapter we shall write φ(D) to denote the function

Dn+a1Dn-1+a2Dn-2+ ... +an-1D+an.

With this notation we may write (i) in the form

φ(D)y = ƒ(x) . (ii)

and, in the case when (x) = 0, we obtain the equation φ(D)y =0 which is known as the reduced

equation of (ii).

2.4. An important property of the reduced equation φ(D)Y=0

If y1, y2,… yn are n solutions of the equation φ(D)y = 0 and A1, A2, ... , An are arbitrary

constants, then A1y1+A2y2+…. +Anyn is also a solution; for

φ(D){A1yl +A2y2+…. + Anyn}

= φ(D)A1yl + φ(D)A2y2+ …. + φ(D)Anyn

= A1φ (D)y1 + A2φ(D)y2 + +Anφ(D)yn=0.

2.5. The solution of the reduced equation φ(D)y=0

We first consider the equation

(D-a)y=0

i.e.

for which e-ax is an integrating factor.

, We have and so

y= Aeax

where A is an arbitrary constant.

When a≠b, the second order equation

{D2-(a+b)D+ab}y=0. (i)

which may be written in the form (D-b){(D-a)y}=0 is clearly

satisfied when

(D-a)y=0

i.e. when y=Aeax.

(i) may also be written in the form (D-a){(D-b)y}=0

 and so is satisfied by y=Bebx.

Hence by § 2.4, (i) is satisfied by y=Aeax+Bebx. This function

which contains two arbitrary constants if a≠b is the general solution of (i).

The nth order equation φ (D)y=0

may be treated in the same way if it is assumed that φ(D) has n distinct linear factors (D -k l), (D-

k2),···' (D -kn), where the k's are real or complex. These factors may be arranged in any order

and hence (24.3) will be satisfied when

(D-k1)y=0, (D-k2)y=O, ... , (D-kn)y=0.

Solving each of these equations, we have

y=A1eklx,y=A2ek2x. ,.,y=Aneknx.

It follows from § 24.4 that

y = A1ek1x + A2ek2x + ... +Aneknx

is the general solution of (24.3) when the k's are all distinct, i.e. when φ(D) has no repeated

factors.

If, however, kl=k2=k3=... =kr, while kr+l, kr+z,···' kn are distinct, the first r terms in (2.4) can be

combined into a single term of the form Ae k1x, the number of arbitrary constants in (24.4) is

reduced to n-r+l and (2.4) does not represent the general solution of

In this case (D - kl) is an r - fold factor of φ (D) .

2.6. The solution of φ(D)y=O when φ(D) has a repeated factor (D-k1)r
 If φ(D) == (D-kl) r(D-kr+1)(D-kr+2)…. (D-kn),

the equation φ(D)y=0 is satisfied when

(D-kl)r =0 (i)

Since one solution of equation (i) is A1ek1x, to find its general solution we put y=ek1xV, where V is

a variable dependent on x.

Then (D-kl)r{ek1xV}=0

i.e. ek1x DrV =0 by

and so DrV =0.

By r successive integrations we obtain

V =Al +A2x+A3x2+ ... +Arxr-1, where the A's are arbitrary constants.

Hence, corresponding to the factor (D - k 1)r in φ(D) , there is the term (A l +A2x+A3x2+ ... +Arxr-
1
)ek1x, containing r arbitrary constants, in the solution of (24.3). The other terms are as found in §

24.5, and the general solution is

v= (Al +A2x+A3x2+ ... +Arxr-1 )eklx+Ar+lekr+1x+ ... +Aneknx In practice the values of k l ,,k2… .. in

any particular case are found by taking y=e kx as a trial solution of the given equation. The

method is illustrated in the following examples.

Example 1

Solve the equation (D2 – 3D- 4)y = O.

This equation is
If y=ekx is a solution

(D2-3D-4)ekx=0

(k2-3k-4)ekx=0

and hence k2-3k-4=0

(ii) is known as the auxiliary equation of (i); its roots are -1 and 4.

Hence, by (24.4), the general solution of (i) is y=Ae-x+Be4x,

where A and B are arbitrary constants.

Example 2

Solve the equation (D2-4)y=0.

The auxiliary equation is k2-4=0; its roots are k=±2. Hence th: general solution is y=Ae2x+Be-2x•

There is an alternative form of solution, for e2x=cosh 2x+sinh 2x an:

e-2x=cosh 2x-sinh 2x.

Hence y=(A+B) cosh 2x+(A-B) sinh 2x

=a cosh 2x+b sinh 2x,

where a and b are arbitrary constants.

Example 3

Solve the equation (D2+9)y=0

The auxiliary equation is k2+9=O; its roots are k=±3i. Hence the general solution is

y=Ae3ix+Be-3ix

But e3ix=cos 3x+i sin 3x and e-3ix=COS 3x-i sin 3x.

Hence y=(A+B)cos 3x+i(A-B) sin 3x (i)

i.e, y=a cos 3x+b sin 3x (ii)

When the roots of the auxiliary equation are purely imaginary, the solution should be

given in trigonometrical form as in (ii) and not in exponential form as in (i).

Useful alternative forms to (ii) are y=R cos (3x+a), y=R sin (3x+a), where R and a are

arbitrary.

Example 4

Solve the equation (D2+4D+13)y=0.

The auxiliary equation is k2+4k+13=0; its roots are k=-2±3i. Hence the general solution

is

y=Ae(-2+3i)x+Be(-2-3i)x

=e-2X(A e3ix + B e-3ix)

=e-2x(a cos 3x+b sin 3x), as in Example 3,

y=Re-2x sin(3x+a).
or

Example 5

Solve the equation (D3+D2-D-1)v=o.

I The auxiliary equation is k3+k2-k-1=0, or (k-1)(k+1)2=0; its i roots are k=-1, -1,1.

Hence, the general solution is y=(A +Bx)e-x+Cex.

Example 6

Solve the equation (D4+8D2+16)y=0.

The auxiliary equation is (k2+4)2=0; its roots are

k=+2i, +2i, -2i, -2i.

Hence, the general solution is

y=(A 1 +B1x)e2ix+ (A2+B2x)e-2ix,

which may be expressed in the form

y=(a1 +b1x) cos 2x+(a2+b2x) sin 2x.

2.7. Solution of the equation φ(D)y=f(x) when f(x)≠0

Let y = u be the general solution of the reduced equation φ(D)y=O i.e. the solution containing n

arbitrary constants; and let us adopt as a trial solution of the equation

 . φ(D)y.= ƒ(x) (i

y=u+v where v, like u, is a function of x.

Then φ(D)(u+v) =ƒ(x)

i.e. φ(D)u+ cp(D)v = ƒ(x).

But φ(D)u=0, :. φ(D)v=ƒ(x) . (ii

Thus v is any function of x which can be found to satisfy (i), and since the n arbitrary constants

required ill the general solution of (i) are contained in u, no arbitrary constant need appear in v.

In other words the general solution (G.S.) of (i) is the sum of two terms u and v: u, the general

solution of the reduced equation φ(D)y=O, which contains n arbitrary constants and is known as th

complementary Junction (C.F.); v, a particular integral (P.L), i.e. function which satisfies (i) and

which contains no arbitrary constant

A particular integral may sometimes be found by inspection, be since, in general, this is not the

case, we go on to consider methods of finding one for equation (i).

From (ii), φ(D) v = ƒ(x)

and we write symbolically f(x)

Our problem is to find a meaning for the operator

2.8. Inverse operators

We define as a function z (if one exists) such that φ(D)z=

According to this definition the operator is the inverse of the operator φ(p) and it can be

shown that can be broken up into factors (which may be taken in any order) or into partial

fractions. It can also be shown that if u and v are functions of x,

Example 12

Solve the equation (D2+6D+8)y=x3e-2x.

The C.F. is Ae-2x+Be-4x.

A P.I. is given by

=
 = )

= - + - )

The general solution is Ae-2x+Be-,x+I e-2x(x4-2x3+3x2-3x). It is usually best to perform the

operation 1/D last.

Example 13

Solve the equation (D2+2D+2)y=50x cos 2x.

The C.F. is e-x(A cos x+B sin x).

Also, x cos 2x is the real part of xe2ix and so to find a P.I. we evaluate

50x =

= (1- +…)x,

= -(i+1)
 =

The real part of this expression, i.e. (11-5x) cos 2x-(2-10x) sin 2x, is a P.I. of the given

equation. Hence the G.S. is

y=e-x(A cos x+B sinx)+{(11-5x) cos 2x-(2-10x) sin 2x}.

2.10. Use of operators in the evaluation of integrals

The theory of operators may be used to evaluate certain types of integrals.

Example 14

∫x4e-2xdx = x4e-2x, where D == d/dx

= e-2x(x4+ 2x3+ 3x2+3 )

the constant of integration being omitted.

2.11 Simple harmonic motion; damped harmonic motion

If a particle moves along the x-axis in such a way that its displace ment from 0 at time t is x,

the velocity of the particle at time t is x and its acceleration is x, where dots denote

differentiation with respect to t.

If the acceleration of the particle is proportional to its displacement from 0 and is directed

towards 0, the differential equation of the motion of the particle is x= -n2x.

This is the differential equation of simple harmonic motion. It solution may be written in

either of the forms

x=a cos nt+b sin nt

or x=R cos (nt+ a),

where a, b, R and a are arbitrary constants.

If, in addition, there is a resistance proportional to the velocity, the motion of the particle is

referred to as damped harmonic motion and the differential equation of the motion is of the

form

+2p +n2x=0,

where the constant p is positive. There are three cases depending on the nature of the roots

of the auxiliary equation.

Example 15

Solve the equation +4 +29x=0.

The roots of the auxiliary equation k2+4k+29=0 are k- 2±5i

and the G.S. is x=e-2t (a cos 5t+b sin 5t),

or x=Re-2t cos (5t+a).

Example 16

Solve the equation +4 +3x=0.

The roots of the auxiliary equation k2+4h+3=0 are k=-1, -3,

and the G.S. is x=Ae-t+Be-3t.

Example 17

Solve the equation

The roots of the and the G.S. is +4 +4x=0.

auxiliary equation k2+4k+4=0 are k=-2, -2,

x=e-2t(A +Bt).

In all three cases, as t→ ∞, x→O, i.e. the motion ultimately dies out and the particle tends to a

position of rest at the origin.

In Example 15 the motion may be roughly described as being oscillatory with constant

period 2 /5 and with decreasing amplitude Re-2t. It can be shown that successive amplitudes

form a decreasing geometrical progression and the motion is said to be slightly damped.
In Example 16 the motion is said to be heavily damped for both x and x tend to zero as

t→∞and so the particle ultimately comes to rest at the origin. If A and B are of opposite sign,

the particle may pass through the origin once before coming to rest, but the motion is not

oscillatory and is known as " dead-beat ".

In Example 17 the motion is said to be critically damped. If A and B are of opposite sign, the

particle may pass through the origin once before coming to rest there, but the motion is not

oscillatory. It is very similar to that of Example 10.

2.12. Forced oscillations

If, in addition to the force which causes simple harmonic motion and the force which causes

damping, the particle is acted on by any other force depending only upon the time, the

differential equation of the motion takes the form

+2k +n2x=f(t)

and the motion is said to be forced.

In the solution of this equation the complementary function 1£ represents the general

solution for free oscillations, i.e. when there is no applied force f(t). The particular integral v

represents the effect of the applied force f(t) on the displacement x.

Since x=u+v and we have already shown that 1£u →0 as t→∞, it follows that x→v as t→∞.

For this reason v is sometimes called the steady state and the part u which dies away is called

the transient.
Example 18

If. + 4 +29x=cos 5t, find x in terms of t and deduce that when t is large the motion of the

particle is approximately simple harmonic and of period 2 /5.

We have shown in Example is that the C.F. is e-2t(a cos 5t+b sin 5t). A P.I. is given by the

real part of

Now = =

and the real part of this expression is (cos 5t + 5 sin 5t).

Hence the G.S. is

x=e-2t (a cos 5t+b sin 5t) + (cos 5t+5 sin 5t).

when t is large, x= (cos 5t+5 sin 5t) approximately, i.e. the motion f the particle is

approximately simple harmonic and of period 2 /5.

2.14 Simultaneous linear equations with constant coefficients

The application of the theory of operators to the solution of simultaneous linear equations with

constant coefficients is demonstrated in the following examples.

Example 21

Solve the equations

+y=sin t+1,

+ x=cos t,

subject to the conditions x=2, y=l, When t=O.

The equations way be written in the form

Dx+y=sin t+ 1………………………………(i)

x+Dy==cos t …………………………………….. (ii)

where D ==- d/dt.

Operating on (i) with D, we have

D2x+Dy=cos t …………………………………….(iii)

and from (ii) and (iii) (D2-1)x=0

which yields x=Aet+Be-t

But, from (i) y=sint+1-Dx

y=sint+ l-Aet+Be-t.

The given conditions, x=2 and Y= 1 when t=0 lead to the equations 2=A +B and O=B-A

so that A =B =1. Hence the particular solutions are

x=et+e-t

y=sin t+ 1-et+e-t.

it should be observed that if y is found from (ii),

Dy=cost_Aet-Be-t

:. y=sint-Aet+Be-t+C .

and substituting from (iv)' and (v) in (i) we find that C = 1.

Again, if y is found by eliminating x between (i) and (ii) ,

(D2-I)y=-2sint-1

:. y=aet+be-t+sin t+ 1 .

Substitution from (iv) and (vi) into (i) or (ii) gives a=-A and b=B, so that there are in

fact only two independent arbitrary constants in the solution of (i) and (ii).
It is useful to note that the number of independent arbitrary constants required in the

solution of the equations

ƒ1 (D)x + F1 (D)y = φ1 (t)

ƒ2(D)x + F2(D)y = φ2(t) ,

where ƒ1,ƒ2 F1 and F2 are polynomials with constant coefficients and D == d/dt, corresponds to

the degree in D of the determinant

To avoid introducing more arbitrary constants than are required, when one of the unknowns

(in this case x) has been determined, the other (in this case y) should if possible be found by

means of a relation which does not involve the derivatives of y. .

Example 22

Solve the simultaneous differential equations

=cos t,

i2=0,

subject to the conditions that i2=i1=O when t=O.

The equations are

(2D+I)i1+Di2=cos t

Di1+(2D+1)i2=O

where D == d/dt.

We eliminate, i1 by operating with D on (i), and with (2D + 1) on (ii)

D(2D+ 1)i1 +D2i2 = -sin t,

D(2D+ l)i1 + (2D+ 1)2i2 =0,

(3D2+4D+1)2i2=sin t.

Whence

The C.F. is Ae-t+Be-kt

A.P.I is the imaginary part of

=- (1 +2i) (cost+isint).

hence a P.I. is - (sin t+2 cos t) and the G.S. is

i2=Ae-t+Be- t- - (Sin 1+2 cos t).

= Be- t - A

The, given equation, are each of the first order so that the expressions for i 1 and i2 should contain

only two arbitrary constants. To avoid introducing further constants we eliminate Di 1 between (i)

and (ii) and obtain i1 in terms of i2 and its derivatives

We multiply (ii) by 2 and subtract from (i) obtaining

Applying the given conditions, We find that

O=B-A + , O=A +B- .

hence A = , B=-

and i1 = (6 cos t+8 sin t-e- -5e-t),

i2= (5e-t-e- t
-4 cos t-2 sin t).

Exercises 24 ( c)
1. Given that x and y are function, of t such that +=3x-y, +=x+y and x=1, y=0

at t=o, show that x-y=e2t

2. Solve the simultaneous equations +ay=a, +ax=y, given that x=0 and y=1

and when t=0

3. A point (x, y) moves in accordance

with the equations

+2y=5et, +-2x=5et.

It is given that x=-l and y=3 when t=0. moves in a straight line. Show that the point

4. Find x and y in terms of t, given that +y=3x, +2x=2y, and that x=0 and y= .

when t=0.