CHAPTER

8
Applications of Differential
Calculus

8.1 Introduction derivative of a function y = f (x), we have

f (x + x) − f (x)
Differential calculus has innumerable applications in the f  (x) = lim
branches of physical sciences, statistics, computer science, x→0 x
engineering, economics, business, actuarial science, demog- where f  (x) represents the slope of tangent to the curve
raphy, and so on. Derivatives are used in engineering and y = f (x) at some point (x, f (x)).
science problems, wherever a problem can be mathemati- Now, if x is very small, then the slope of tangent is
cally modelled and an optimal solution is desired. approximately the same as the slope of the secant through
Sometimes, a graphical representation helps for better (x, f (x)) and (x + x, f (x + x)). Thus
understanding of the behaviour of a function. For exam- f (x + x) − f (x)
f  (x) ≈
ple, an engineer or a scientist finds a suitable mathemati- x
cal model to represent a physical system. While examining Replacing x with differential d x, we have
the functions used in the mathematical model, the local
maxima, local minima, inflection points, and the changes f  (x) d x ≈ f (x +d x)− f (x) or f (x +d x) ≈ f (x)+ f  (x) d x
in inflection have physical interpretations and these allow This gives an approximation formula for the functional
the engineer or the scientist to account for the behaviour of value of f at a point x + d x (close to x) in its domain,
the physical system. Curve tracing, the process for graph- when the values of the function and its derivative at x are
ical representation of a function also involves differential known.
calculus. Writing y for f (x), we have
In this chapter, some common applications of differen-
tial calculus, namely approximation of a function at some dy = f (x + d x) − f (x) ≈ f  (x) d x
point, estimate of errors in the approximation, measuring Example 8.1 Find dy if y = x 2 + 7x + 2.
arc-length, finding curvature, curve tracing, and optimiza-
Solution Let y = f (x) = x 2 + 7x + 2.
tion of the value of a function with or without constraints
Then, f  (x) = 2x + 7.
are considered.
Therefore, dy ≈ f  (x)d x = (2x + 7)d x.
Example 8.2 Find dy if y = cos x + sin 2x.
8.2 Approximations and Errors
Solution Let y = f (x) = cos x + sin 2x.
From the definition of derivative, an approximate value of Then, dy ≈ f  (x)d x = (2 cos 2x − sin x)d x.
a function at a point can be computed and the error in
this approximation can also be estimated. When a bet- 8.2.1 Approximations
ter approximation is required, a function may be approxi-
mated by a polynomial using Lagrange series or Maclaurin Let us now illustrate with some examples how the value of
series. a function at a given point in its domain can be approxi-
Let us begin by finding the formula that ultimately en- mated.
ables the approximation of the value of a function at a Example 8.3 Find an approximate value of 2x 3 + 7x + 1 at
given point in its domain. Recalling the definition for the x = 2.001.
 Applications of Differential Calculus 169

Solution Let f (x) = 2x 3 + 7x + 1, then f  (x) = 6x 2 + 7. Example 8.7 Using Taylor series expansion, calculate an
For the evaluation of f (x), a convenient value of x close to approximate value of cos 58◦ . Take π = 3.141593.
2.001 is x = 2.
Solution We write f (x) = cos x and express f (x) as Taylor
We have f (2) = 31. π
series about x = , which is very close to 58◦ , as the values
Also, as x increases from x = 2 to x = 2.001, we have 3 π
d x = 0.001. of cos x and its derivatives are known for x = .
3
Now, f  (x) = 6x 2 + 7, so, f  (2) = 31. Differentiating f (x) successively, we have
Hence, dy ≈ f  (x)d x = 31 × 0.001 = 0.031.
f  (x) = − sin x, f (2) (x) = − cos x, f (3) (x) = sin x
Thus, for x = 2.001, an approximate value of f (x) is π
31 + 0.031 = 31.031. For x0 = , we have
3
1 √ √
Example 8.4 Find an approximate value of (8.05) 3 . 1 3 1 3
 (2) (3)
1 1 2 f (x0 ) = , f (x0 ) = − , f (x0 ) = − , f (x0 ) =
Solution Let f (x) = x 3 , then f  (x) = x − 3 . 2 2 2 2
3 Thus, expanding f (x) in Taylor series about x0 , we can write
Take a = 8 and h = 0.05.
√ 
π
1
Then, f (a) = f (8) = 2 and f  (a) = = 0.083. 1 3
12 f (x) = cos x ≈ − x−
Thus, by the approximation formula 2 2 3
 π 2  π 3
1
(8.05) 3 = f (8.05) ≈ f (8) + 0.05 × f  (8) x − √ x −
1 3 3 3
= 2 + 0.05 × 0.083 − +
2 2! 2 3!
= 2.00415 58π
Example 8.5 Find an approximate value of tan−1 (0.99). Putting x = 58◦ = , so that
180
Solution Let f (x) = tan−1 x. π 2π 3.141593
x− =− =− = −0.0349
Then 3 180 90
1 we have
f  (x) = 1.732 1
1 + x2 f (x) = cos x ≈ 0.5 − (−0.0349) − (−0.0349)2
Take a = 1 and h = −0.01. 2 4
1.732
Then + (−0.0349)3
π 12
f (a) = f (1) = tan−1 (1) = radians and 0.00122
4 ≈ 0.5 + 0.866 × 0.0349 −
1 4
f  (1) == 0.5 − 0.1441 × 0.00004
2
By the approximation formula f (a + h) ≈ f (a) + h f  (a), we ≈ 0.5 + 0.0302 − 0.000305 − 0.000006 ≈ 0.5299
have Thus, for x = 58◦ , cos x ≈ 0.5299.
tan−1 (0.99) ≈ f (1) − 0.01 × f  (1) Example 8.8 Using Maclaurin series expansion, calculate
π √
= − 0.01 × 0.5 ≈ 0.7804 radians. an approximate value of 15.
4 
Example 8.6 Find an approximate value of sin 31◦ , assum- √ √ 1
Solution We write 15 = 16 − 1 = 4 1 − .
ing that 1◦ = 0.0175 radians and cos 30◦ = 0.8660. 16
1 1
Solution Let f (x) = sin x, then f  (x) = cos x. Consider the function f (x) = (1 − x) 2 , where x = .
16
π Then
Take a = 30◦ = radians and h = 1◦ = 0.0175 radians. 1 1
Then
6 f  (x) = − (1 − x)− 2
2
f (a) = f (30◦ ) = sin 30◦ = 0.5 and 1 3
f (2) (x) = − (1 − x)− 2
f  (a) = f  (30◦ ) = cos 30◦ = 0.8660 4
3 5
Using the approximation formula f (a + h) ≈ f (a) + h f  (a), f (3) (x) = − (1 − x)− 2
8
we get For x = 0, we have
sin 31◦ ≈ f (30◦ ) + 0.0175 × f  (30◦ ) 1 1 3
f (0) = 1, f  (0) = − , f (2) (0) = − , f (3) (0) = −
or sin 31◦ ≈ 0.5000 + 0.0175 × 0.8660 = 0.5152. 2 4 8
Thus, expanding in Maclaurin series, we can write
Approximations using Taylor and Maclaurin Series 1 1 1 x2 3 x3
f (x) = (1 − x) 2 ≈ 1 − x− −
The following examples demonstrate the techniques for find- 2 4 2! 8 3!
ing an approximate values of functions using Taylor and x x2 x3
= 1− − −
Maclaurin series expansion. 2 8 16
170 Engineering Mathematics

1 8.2.2 Errors
Putting x = , we have
16
 Let y = f (x) be a differentiable function of x and d x an
15 1 1 1 error in x, then the approximate error in y is dy ≈ f  (x) d x.
≈1− − − 4
16 32 256 × 8 16 Approximate relative error in y is given by
Thus, we get dy f  (x) d x
=
√ 1 1 1 y f (x)
15 ≈ 4 − − −
8 512 4 × 163 Approximate percentage error in y is given by
≈ 4 − 0.125 − 0.00195 − 0.00006 ≈ 3.87299
dy f  (x) d x
100 × = 100 ×
Example 8.9 Using Taylor series expansion, calculate an y f (x)
approximate value of e x cos x for x =42◦ . Assume 1◦ = 0.0175
π π π Example 8.10 The radius of a sphere is measured as 15
radian, e 4 = 1.1934, and e 4 cos = 1.551.
4 cm with an error of 0.01 cm. Find (a) the approximate error,
Solution We write f (x) = e x cos x and express f (x) as Tay- (b) relative error, and (c) percentage error in calculating its
lor series. volume.
Using Result 6.10 of successive differentiation, we have
√ n  nπ 
Solution If r is the radius of the sphere, then its volume is
f (n) (x) = 2 e x cos x + for n = 1, 2, 3, . . . 4
V = πr 3 .
4 3
π dV
Thus, taking x0 = 45◦ = , we have Therefore, = 4πr 2 .
4 dr
π  dV
π Approximate error in volume, d V = dr = 4πr 2 dr .
f (x0 ) = e 4 cos dr
4 4
√ π π  When r = 15 cm and dr = 0.01 cm, we have V = π(15)3
 3
f (x0 ) = 2e 4 cos =0
2 cubic cm and d V = 4π(15)2 (0.01) cubic cm.
  π 
π 3π π Thus, the relative error is given by
f (2) (0) = 2e 4 cos = −2e 4 sin
4 4 dV 4π(15)2 (0.01) 3 × 0.01
  = = = 0.002
√ π 4π √ π V 4 15
f (3) (0) = 2 2e 4 cos = −2 2e 4 π(15)3
4 3
π Percentage error is given by
Thus, expanding f (x) in Taylor series about x0 = , we can
4 dV
write 100 × = 100 × 0.002 = 0.2 .
V
f (x) = e x cos x
 π 2 Example 8.11 The time T taken for a completeoscillation
π
π  π
 π  x − l
≈ e 4 cos − 2e 4 sin 4 of a simple pendulum of length l is given by T = 2π , where
4 4 2! g
 π 3
g is a constant. If there is an error of 1% in the value of l, find
π
x− the approximate error and percentage error in the calculated
−2e 4 4
3! value of T .
π  π   
π π π 2 l π
= e 4 cos − e 4 cos x− Solution We have T (l) = 2π , and hence, T  (l) = √ .
4 4 4 g gl
√
2 π  π 3 Now, percentage error in l is 100 ×
dl
= 1, and hence,
− e4 x − l
3 4 l
π dl = .
Putting x = 42◦ so that x − = 3 × 0.0175 = 0.0525 radian, 100
4 Therefore, approximate error in T is
we have

 π l π l
e x cos x ≈ 1.551 − 1.551 × (0.0525)2 dT = T (l) dl = √ =
1.414 × 1.1934 g l 100 100 g
− (0.0525)3
3 Percentage error in T is
≈ 1.551 − 1.551 × 0.00276 − 0.000057 
dT π l 1
≈ 1.551 − 0.00427 − 0.000057 ≈ 1.547 100 × = 100 ×  = 0.5.
T 100 g l
2π
Thus, for x = 42◦ , e x cos x ≈ 1.547. g
 Applications of Differential Calculus 171

Section Review 8.1

2
1. Find an approximate value of the function y = (x − 1) 5 the sports wear is priced at x per unit. Suppose the
for x = 1.05. current value is 500 per unit. If the price is raised to
2. Find an approximate value of cos 61◦ , assuming that 510 per unit, what will be the effect on the sale of the
sin 60◦ = 0.8660 and 1◦ = 0.0175 radians. item?
3. Find an approximate value for each of the following. 5. Calculate the error and percentage of error for the
√ 3 approximation in Question 1.
(a) 15.7 (b) (3.9) 2 (c) log(0.98) (d) sin(0.1135)
4. The demand equation of a sports wear is given by 6. Find an approximate value of sin 31◦ , assuming that
2250 cos 60◦ = 0.8660 and 1◦ = 0.0175 radians. Calculate also
s= , where s represents the monthly sales, when the error and percentage of error.
x

8.3 Arc-length Substituting these in Eqn (8.3), we get

In this section, formulae for finding the arc-length of a plane f  cos θ − f sin θ
tan ψ = (8.6)
curve (described in two-dimensional co-ordinate plane) be- f  sin θ − f cos θ
tween two given points on it are derived. Now, from the triangle OPT (see Fig. 8.1), we have
θ + φ = ψ or φ = ψ − θ .
Therefore, we get
8.3.1 Angle between Tangent and Radial Vector
tan ψ − tan θ
Let the equation of the curve in polar form be r = f (θ ), tan φ =
1 + tan ψ tan θ
and P(r, θ ) be any point on the curve. Taking the pole as sin θ
the origin and the initial line as the x-axis, if the Cartesian ( f  sin θ + f cos θ ) /( f  cos θ − f sin θ) −
= cos θ
co-ordinates of the point are (x, y), then sin θ
1 + {( f  sin θ + f cos θ )/( f  cos θ − f sin θ)}
cos θ
x = r cos θ = f (θ ) cos θ (8.1) ( f  sin θ cos θ + f cos2 θ ) − ( f  sin θ cos θ − f sin2 θ)
=
( f  cos2 θ − f sin θ cos θ ) + ( f  sin2 θ + f sin θ cos θ)
y = r sin θ = f (θ ) sin θ (8.2)
f r dθ
=  = =r .
In Fig. 8.1, let P T be the tangent to the curve at P (r, θ ) f dr dr
and ψ be the angle that tangent makes with the positive dθ
x-axis. Let φ be the angle between the radius vector O P Thus
and the tangent P T , then dθ
tan φ = r (8.7)
dy dy/dθ dr
tan ψ = = (8.3) Illustration 8.1 Consider the polar curve r = 2 sin θ ,
dx d x/dθ
which is a straight line parallel to the x-axis and at a
Differentiating Eqn (8.1) and Eqn (8.2) w.r.t. θ, we have
distance of 2 units from it.
dx df Differentiating r = 2 sin θ w.r.t. r , we have
= cos θ − f sin θ = f  cos θ − f sin θ (8.4)
dθ dθ dθ dθ 1
1 = 2 cos θ or =
dy df dr dr 2 cos θ
= sin θ + f cos θ = f  sin θ + f cos θ (8.5) At any point (r, θ) on the curve, we have
dθ dθ
dθ 2 sin θ
tan φ = r = = tan θ
dr 2 cos θ
Thus, at any point (r, θ) on the curve, φ = θ .

 Note If two curves intersect a point P(r, θ ), the angle

φ of intersection of the two curves is defined to be the
angle (0–π /2) between the tangents to the curves at the
point. Let φ1 be the angle between the radius vector and
the tangent to one curve and φ2 be the angle between the
x radius vector and the tangent to the second curve. Then the
O T
angle of intersection α of the curves is either α = φ2 − φ1
Fig. 8.1 Tangent and radial vector of a curve or α = φ1 − φ2 .
172 Engineering Mathematics

8.3.2 Differential of Arc-length we have

  2
For a continuous curve, its arc-length up to a point on the ds dy
curve, measured from a fixed point on it, is the distance = 1+
dx dx
between the two points measured along the curve. It is,
The above equation gives the rate of change of arc-length
thus, a function of the co-ordinates of the point up to which
w.r.t. x.
it is measured. If the point is given a small shift along the
The differential of arc-length is given by
curve, then there may be a small change in the arc-length 
also. This change in arc-length corresponding to a small  2 
dy
shift of the point is called the differential of arc-length at ds = 1 + d x = (d x)2 + (dy)2 (8.8)
dx
the point. The expressions for the differential of arc-length
of a continuous curve are derived in the following text for  Notes
different co-ordinate systems. 1. If the equation of the curve in Cartesian co-ordinates
is x = g(y), the rate of change of arc-length w.r.t. y is
Cartesian Form given by
  2
Let the equation of the curve in Cartesian co-ordinates be
ds dx
y = f (x). Let A(x0 , y0 ) be a fixed point on the curve from = 1+
dy dy
which the arc-length is measured. In Fig. 8.2, let P(x, y)
be any point on the curve and s be the length of the arc However, the differential of arc-length remains the same,
A P. Let Q(x + x, y + y) be a neighbouring point of P as
  2
and the length of the arc AQ be s + s. 
dx
Then, we have P Q 2 = (x)2 + (y)2 . ds = 1 + dy = (d x)2 + (dy)2
dy
Dividing both sides by (x)2 and rewriting, we have
          2. If the tangent to the curve y = f (x) at P(x, y) makes
PQ 2 y 2 P Q 2 s 2 y 2 dy
=1+ or =1+ an angle ψ with the x-axis, then = tan ψ.
x x s x x dx
Therefore
Making Q → P so that x → 0, we have   2 
s ds
= 1+
dy
= 1 + tan2 ψ = sec ψ
P Q → 0 and lim =1 dx dx
Q→P PQ
Thus, taking limits as x → 0, we have dx
and hence, cos ψ = .
      ds
PQ 2 s 2 y 2 Again
lim lim = 1 + lim
x→0 s x→0 x x→0 x ds ds d x 1
 2  2 = = sec ψ = cosecψ
ds dy dy d x dy tan ψ
or =1+
dx dx dy
and hence, sin ψ = .
Assuming that arc-length s increases as x increases so that ds
ds Example 8.12 Determine the length of the curve defined
>0 π
dx by y = log (sec x) between 0 ≤ x ≤ .
4
Solution Here, we have
Y  2
dy sec x tan x dy
= = tan x and = tan2 x
Q dx sec x dx
⌬s So, we get
⌬y 
P  2
N dy
⌬x 1+ = | sec x| = sec x
A dx
π
as sec x is positive for 0 ≤ x ≤ .
␺ 4
Therefore, the arc-length is
O T X π
4 π √ 
sec xd x = log sec x + tan x 4
0
= log 2+1 .
Fig. 8.2 Differential arc-length in Cartesian co-ordinates 0
 Applications of Differential Calculus 173

Illustration 8.2 Let us compute the arc-length of As Q → P,

2 3
x = (y − 1) 2 P Q → 0 and θ → 0
3
between 1 ≤ y ≤ 4. and also
The equation of the curve is in the form x = g(y). So, it s length of arc PQ sin θ
dx = → 1 and →1
is easy to compute . PQ length of chord PQ θ
dy
We have So, making Q → P and taking limits, we get
dx 1  P Q s 2  sin θ 2
= (y − 1) 2 lim 2
=r lim
dy Q→P s θ θ → 0 θ
So, we get ⎧  θ  ⎫2
 ⎪
⎨ sin  θ  ⎪
⎬
 2 r 2
dx √ + lim + 2r sin
1+ = y θ → 0 ⎪
⎩ θ θ 2 ⎪
⎭
dy
Thus, the arc-length is
  2 2  2
 2 ds dr dr
4
dx 4
√ 2 3 4 2 14 i.e., = r ×1 +
2
+ 2r × 1 × 0 =r +
2
1+ dy = ydy = y 2 = [8 − 1] = dθ dθ dθ
1 dy 1 3 1 3 3
Therefore
  2
Polar Form ds dr
Let the equation of the curve in polar co-ordinates be = r2 +
dθ dθ
r = f (θ ). Let A(r0 , θ0 ) be a fixed point on the curve from   2
which arc-length is measured. In Fig. 8.3, let P(r, θ ) be dr
ds = r +
2 dθ (8.9)
any point on the curve and Q(r + r, θ + θ ) a neigh- dθ
bouring point of P. Let s be the length of the arc AP and A point A (r0 , θ0 ) can always be chosen such that r increases
s + s be that of the arc AQ. as θ increases, therefore, in Eqn (8.9) positive sign of the
Suppose PN is drawn perpendicular to OQ from P. radical is taken.
Illustration 8.3 Let us compute the length of the loga-
Q(r⫹⌬r, ␪⫹⌬␪) rithmic spiral r = eθ for the first rotation, i.e., from θ = 0
N to θ = 2π .
The arc-length is
  2 2π 
P(r, ␪) 2π
dr
r +
2 dθ = e2θ + e2θ
␾ 0 dθ 0
⌬␪ 2π √
␺
␪ = 2eθ dθ
O T X 0
√  2π √
Fig. 8.3 Differential arc in polar co-ordinates = 2eθ = 2 e2π − 1
0

Now,  OPN is right-angled at N .  Notes

Therefore, O N = OP cos θ = r cos θ, P N = O P 1. The rate of change of arc-length w.r.t. that of r may be
sin θ = r sin θ , OQ = r + r ,and NQ = OQ − ON = derived as follows:

θ  2
r + r − r cos θ = r + 2r sin2 . ds ds dθ dr dθ
2 = = r2 +
dr dθ dr dθ dr
From  PNQ which is right-angled at N , we have  
 2  2   2  2  2
θ dθ dr dθ dθ
P Q = P N + N Q = r sin θ + r + 2r sin
2 2 2 2 2 2 = r2 + = 1 + r2
2 dr dθ dr dr

Dividing both sides by θ 2 and rearranging, we get ds

To find , it is better to express the polar equation of
⎧   ⎫2 dθ
⎪ θ a curve in the form θ = g(r ).
    ⎪  ⎪
⎪
PQ 2 2 sin θ 2 ⎨r θ ⎬
sin 2. In Fig. 8.3, as Q → P, OQ coincides with the line
2
=r + + 2r sin
θ θ ⎪
⎪ θ θ 2 ⎪⎪
OP and chord PQ coincides with the tangent TP at P.
⎩ ⎭ Therefore, the angle φ between the tangent PT at P and
174 Engineering Mathematics

the radial vector OP is the limiting value of the angle Example 8.14 For a curve, if f  (θ ) = x sin θ + y cos θ and
 PQN as Q → P.
f  (θ ) = x cos θ − y sin θ , show that its arc-length s is given by
Thus s = f (θ ) + f  (θ) + c, where c is a constant.
PN r θ dθ
tan φ = lim = lim =r Solution Differentiating f  (θ) = x sin θ + y cos θ w.r.t. θ , we
Q→ P QN r → 0 r dr
Now have
  
r dθ 2 
dx dy
ds f  (θ ) = sin θ + x cos θ + cos θ − y sin θ
= 1+ = 1 + tan2 φ = sec φ dθ dθ
dr dr
Thus dx dy
dr dr dθ dθ or sin θ + cos θ = f  (θ) − f  (θ) = 0 (8.10)
cos φ = and sin φ = cos φ tan φ = r =r dθ dθ
ds ds dr ds
Similarly, differentiating f  (θ ) = x cos θ − y sin θ w.r.t. θ, we
Example 8.13 Find the length of the arc in one period have
(t = 0 to t = 2π ) of the cycloid x = a(t − sin t), y = a(1−cos t). dx dy
2π  f (3) (θ) = cos θ − x sin θ − sin θ − y cos θ
dθ dθ
Solution Arc-length = (d x)2 + (dy)2
t=0
 dx dy
2π  2  2 or cos θ − sin θ = f (3) (θ ) + f  (θ ) (8.11)
dx dy dθ dθ
= + dt
0 dt dt Squaring Eqn (8.10) and Eqn (8.11) and then adding, we get

2π  2  2
= a 2 (1 − cos t)2 + a 2 sin2 t dt dx dy  2
0 + = f  (θ) + f (3) (θ)
dθ dθ
2π √
=a 2 − 2 cos t dt Thus, the arc-length is given by
0

2π
t  2  2
= 2a sin dt dx dy
0 2 s= + dθ
 dθ dθ
π
t  
= 4a sin u du putting u =
0 2 = f  (θ ) + f (3) (θ ) dθ
π
= 4a (− cos u)|0 = 8a.
= f (θ) + f  (θ) + c, where c is a constant.

Section Review 8.2

1 1
1. Find the length of the arc of the curve y = x 3 + x −1 5. Find the lengths of the arc of the following polar curves.
6 2 (a) r = cos θ, 0 ≤ θ ≤ 2π
between the ordinates x = 1 and x = 2.
π
2. Find the length of the arc of the curve y = log sin x (b) r = 1 + sin θ, 0 ≤ θ ≤
π π 2
between
6
and .
3 (c) r = e−θ , 0 ≤ θ ≤ 1
π
3. Calculate the arc-length of y = x 2 between x = 0 and (d) r = sec θ, 0 ≤ θ ≤
x = 2. 3 π
3 (e) r = 1 − sin θ, 0 ≤ θ ≤
4. Calculate the arc-length of the curve y = x 2 between 2
6. Calculate length of the parametric curve x(t) = t −
x = 0 and x = 1. π
sin t, y(t) = 1 − cos t for 0 ≤ t ≤ .
4

8.4 Curvature
The shape of a curve depends largely on the rate at which Consider a plane curve ABC as shown in Fig. 8.4. Let
the direction of the tangent changes as the point of contact the tangents at points A and B make angles α and α + α,
moves along the curve. If this rate is high, then the bending respectively, with the positive direction of x-axis.
of the curve is more at the point. This rate of change of Then the average curvature of the arc AB is given by
direction with respect to its arc-length at a point is called
α α
curvature of the curve at that point and is denoted by κ. =
length of arc AB s
For example, a straight line is a curve with curvature equal
to zero and a circle has a fixed curvature. where s = length of arc AB.
 Applications of Differential Calculus 175

C Hence, at that point, the curvature is

4a 2 4a 2 1
K =− 3 = − √ =− √ .
(4a + 4a )
2 2 2 16 2a 3 4 2a

8.4.2 Parametric Form

B ⌬␣
Let the curve be defined in parametric form as x = x(t),
⌬s
y = y(t), t being the parameter.
A
Then, we have

dy dy dx y
y1 = = = 
dx dt dt x
␣ ␣⫹⌬␣
d x dy
O X where x  = and y  = .
dt dt
   
d2 y d dy d dy dx
Fig. 8.4 Curvature at a point y2 = = =
dx2 dx dx dt d x dt
To get the curvature of the curve at A, make B approach
x  y  − x  y  1 x  y  − x  y 
the point A along the curve and take the limit of the above = =
(x  )2 x (x  )3
expression.
Thus, the curvature of the curve at A is given by Substituting in Eqn (8.12) the values of y1 and y2 , we have
   2 32
α α dα y2 x  y  − x  y  y
K = lim = lim = K =
B → A Length of arc AB s → 0 s ds 3 = 1+
1 + y2 2 (x )
 3
x
1
Let us derive more specific expressions for the curvature of
x  y  − x  y 
a curve in different co-ordinate systems. = ! "3 (8.13)
(x  )2 + (y  )2 2
8.4.1 Cartesian Form
Illustration 8.5 For the cardioid x = 2 cos t − cos 2t,
Let the equation of the curve in Cartesian form be y y = 2 sin t − sin 2t, let us find the curvature at a point
= f (x). π
dy where t = .
Then tan α = = y1 or f  (x) so that α = tan−1 y1 , 3
dx The equations of the curve are x = 2 cos t − cos 2t,
dα 1 d y2 y = 2 sin t − sin 2t. Therefore, we have
and hence, = (y1 ) = .
dx 1 + y12 d x 1 + y12 dx d2x
 x = = −2 sin t + 2 sin 2t, x  = 2 = −2 cos t + 4 cos 2t
ds dt dt
We know = 1 + y12 .
dx dy d2 y
dα dα d x y2 1 y = = 2 cos t − 2 cos 2t and y  = 2 = − 2 sin t + 4 sin 2t
Therefore, κ = = =  . dt dt
ds d x ds 1 + y12
1 + y12 π
For t = , we have
Thus, we obtain 3
√  √     
dα y2 f  (x)  3 3 1 1
κ= = 3 =  3 (8.12) x = −2 +2 = 0, x  = −2 +4 − = −3
ds 1 + y2 2 1 + { f  (x)}2 2 2 2 2 2
1     √  #√ $
 1 1  3 3 √
 Note If the equation of curve is x = f (y), the curva- y =2
2
−2 −
2
= 2, y = −2
2
+4
2
= 3
ture is given by
  2  32 Substituting these values in Eqn (8.13), we have the
dα d2x dx f  (y) curvature
κ= = 2
1+ or κ =  3 √
ds dy dy 1 + { f  (y)}2 2 0 × 3 − 2 × (−3) 6 3
κ= 3 = 3 =
Illustration 8.4 Consider the parabola y 2 = 4ax and let 0 + 22 2 2 4
us find its curvature at the upper end of its latus rectum.
We have 8.4.3 Polar Form
dy 2a d2 y 2a dy 4a 2 Suppose the equation of the curve in polar form is r = f (θ ),
= and 2
=− 2 =− 3
dx y dx y dx y where r and θ are polar co-ordinates.
Substituting in Eqn (8.12), the curvature at any point is The Cartesian co-ordinates (x, y) of any point having
4a 2 polar co-ordinates (r, θ ) can be expressed as
κ=− 3 .
(y 2 + 4a 2 ) 2 x = r cos θ = f (θ ) cos θ
The upper end of the latus rectum is (a, 2a). y = r sin θ = f (θ ) sin θ
176 Engineering Mathematics

Therefore, we obtain 8.4.4 Implicit Form

dx If the equation of a curve is given in implicit form
= r1 cos θ − r sin θ
dθ f (x, y) = 0, then we have
dy
= r1 sin θ + r cos θ dy f x (x, y) fx
dθ =− =−
dx f y (x, y) fy
Thus where f x = f x (x, y), f y = f y (x, y); assuming that

dy dy dx r1 sin θ + r cos θ f y (x, y) = 0.
= =
dx dθ dθ r1 cos θ − r sin θ Now
     
d2 y d dy d dy dθ dy dy
and = = fy fx x + fx y − f x f yx + f yy
dx2 dx dx dθ dx dx d2 y dx dx
 = −  2
d r1 sin θ + r cos θ 1 dx2 fy
= % & % &
dθ r1 cos θ − r sin θ r1 cos θ − r sin θ
 − f y f x x − f x y f y + f x f yx − f yy ff xy
fx
(r1 cos θ − r sin θ ) (r2 sin θ + 2r1 cos θ − r sin θ) =  2
− (r1 cos θ − r sin θ ) (r2 cos θ − 2r1 sin θ − r cos θ) fy
= ! 2 " ! "
(r1 cos θ − r sin θ )2 − f y f x x + f x f y f x y + f x f y f yx − f x2 f yy
1 r 2 + 2r12 − r r2
=  3
× = fy
r1 cos θ − r sin θ (r1 cos θ − r sin θ )3
2 f x f y f x y − f y2 f x x − f x2 f yy
=  3 as f yx = f x y
Thus, the curvature is given by fy
y2
κ= 3
dy d2 y
1+ y12 2 Substituting and in Eqn (8.12), we have
dx dx2
  2 3

r 2 + 2r12 − rr2 r1 sin θ + r cos θ

2 2 f x f y f x y − f y2 f x x − f x2 f yy
= 1+ κ=  3 (8.15)
(r1 cos θ − r sin θ )3 r1 cos θ − r sin θ f x2 + f y2 2
Illustration 8.7 Let us find the curvature of the curve
i.e.,
defined by f (x, y) = x 2 y + y 3 + 2 = 0 at the point (1, −1).
r 2 + 2r12 − r r2 Here, we have f x = 2x y, f x x = 2y, f y = x 2 + 3y 2 ,
κ= ! "3 (8.14)
r 2 + r12 2 f x y = 2x, and f yy = 6y.
At point (1, −1), we have f x = −2, f x x = −2,
dr d 2r
where r1 = and r2 = . f y = 1 + 3 = 4, f x y = 2, and f yy = −6.
dθ dθ 2
Therefore, the curvature is given by
Illustration 8.6 Consider the logarithmic spiral r = eaθ
2 f x f y f x y − f y2 f x x − f x2 f yy
and find the curvature at any point on it. κ=  3
dr d 2r f x2 + f y2 2
We have = aeaθ = ar and = a 2 eaθ = a 2r .
dθ dθ 2 2(−2)(4)(2) − (4)2 (−2) − (−2)2 (−6)
Substituting in Eqn (8.14), we have =  3
(−2)2 + (4)2 2
r 2 + 2(ar )2 − r (a 2r ) 1 √
κ= ! " 32 = √ −32 + 32 + 24 24 3 5
r 2 + (ar )2 r 1 + a2 = 3 = √ =
(4 + 16) 2 40 5 25

Section Review 8.3

1. Find the curvature of the curve x 5 + 4x y 3 − 3y 5 = 0 at 5. Find the curvature of the curve r = 2+ sin 3θ at the point
the point (1, −1). (2, 0).
2. Find the curvature of the curve y 3 = a 2 x at any point 6. Find the curvature of the curve r = 3 cos θ at the point
(a, a) on it. (3, 0).
3. Find the curvature of the parabola y 2 = 4x at the vertex. 7. Find the curvature of the curve defined by the parametric
4. 
Find the curvature of the curve r = 1 − sin θ at the point equations x = 3t, y = t 2 + 2 for any t.
1 π 8. Find the curvature at any point of the catenary
, .
2 6 a x x

y= e a + e− a
2
 Applications of Differential Calculus 177

8.5 Radius of Curvature dr d 2r

where r1 = and r2 = .
dθ dθ 2
If P is any point on a curve, the radius of curvature of the
Illustration 8.10 Consider again the logarithmic spiral
curve at P is defined to be the reciprocal of the curvature
r = eaθ of Illustration 8.6 and find the radius of curvature
at that point. The radius of curvature is generally denoted
1 of at any point (r, θ ).
by R, and thus, R = . Generally, the numerical value is 1
κ For this curve, we have κ = √ (see Illustration 8.6).
taken. r 1 + a2
Therefore, the radius of curvature of the curve at any
The radius of curvature for the different forms of the 1 
equation of a curve can be expressed as follows. point (r, θ ) is R = = r 1 + a 2 .
κ
8.5.1 Cartesian Form
8.5.4 Implicit Form
If the equation of a curve is y = f (x), then the radius of
If the equation of a curve in implicit form is f (x, y) = 0,
curvature is given by
 then
 2  32  2 3
 2 3
1 dy d y 1 + y12 2 1 f x + f y2 2
R = = 1+ = R= =
κ dx dx2 y2 κ 2 f x f y f x y − f y2 f x x − f x2 f yy
If the equation of the curve is x = f (y), then the radius of Illustration 8.11 Consider again the equation f (x, y) =
curvature is given by x 2 y + y 3 + 2 = 0 of Illustration 8.7 and find the radius of
  2  32  2
1 dx d x curvature of the curve at the point (1, −1).
R = = 1+ 1
κ dy dy 2 For this curve, we have κ = (see Illustration 8.7).
2
Therefore, the radius of curvature of the curve at the
Illustration 8.8 Consider the parabola y 2 = 4ax of 1
Illustration 8.4 and let us find its radius of curvature at point (1, −1) is R = = 2.
κ
the upper end of its latus rectum.
At the upper end of the latus rectum, (a, 2a), the cur-
1 8.5.5 Pedal Form
vature is given by κ = − √ (see Illustration 8.4).
4 2a Let P(r, θ ) be any point on a curve with polar equation
Therefore, the radius of curvature at the point is f (r, θ ). Let P T make an angle ψ with the initial line OX,
1 √ O being the pole (see Fig. 8.5).
R = = 4 2a .
κ Draw ON perpendicular to the tangent PT and let
p = ON.
8.5.2 Parametric Form Let φ be the angle between the tangent and the radial
If a curve is defined in parametric form x = x(t), y = y(t), vector, then
t being the parameter, then ON p
sin φ = =
!  2 "3 OP r
1 (x ) + (y  )2 2 Therefore
R= =
κ x  y  − x  y  p = r sin φ (8.16)
Illustration 8.9 Consider again the cardioid x = 2 cos t − Equation (8.16), which relates the radial vector at any
cos 2t, y = 2 sin t − sin 2t of Illustration 8.5. Let us find the point on the curve to the perpendicular distance of the
π
radius of curvature at the point t = .
3
3
For this cardioid, we have κ = (see Illustration 8.5).
4
Therefore, the radius of curvature at the point is
1 4 P(r, ␪)
R= = . r
κ 3 ␾
␪ ␺
8.5.3 Polar Form O T
P
If the equation of a curve is r = f (θ ), where r and θ are
the usual polar co-ordinates, then
! 2 "3 N
1 r + r12 2
R= = 2 Fig. 8.5 Pedal curve
κ r + 2r12 − r r2
178 Engineering Mathematics

tangent at that point from the pole, is called a pedal equa- a

When r = √ , we have
tion and the curve is called a pedal curve. 2
dθ 1 dr √   √
We know tan φ = r , and hence, cot φ = . 2a 5 4 5
dr r dθ √ a √ a5 √
Now, cosec φ = 1 + cot φ.
2 2 dr
=
5 4
= 5
2 2
√ =
5 5
Therefore dp a 2a 2 a 3 a a5 2 6
  a4 √ − 2 √ √ −
2 5 2 2 2 5
r2 1 dr 2
=1+ 2 a
p2 r dθ Thus, the radius of curvature when r = √ is
2
 2 ' ( √ √
1 1 1 dr dr a 5 5 5 5
or = 2+ 4 (8.17) r = √ = √ a.
p 2 r r dθ dp r = √a 2 6 6 2
2

Equation (8.17) is another form of the pedal equation of a

curve.
8.5.6 Radius of Curvature at the Origin or Pole
Radius of Curvature for Pedal Curve
Let us look at the radius of curvature of a curve at the
We have p = r sin φ. origin using the following forms for the equation of the
dp dφ dr dφ dθ given curve.
= cos φ + sin φ = r +r
dr dr ds dr ds
Cartesian Form
dr dθ
since cos φ = and sin φ = r . Let a curve y = f (x) pass through the origin.
ds ds
From Fig. 8.5, we have ψ = θ + φ. One of the following three cases can occur:
Hence Case 1 The 'x-axis
( is tangent to the curve at the origin.
dp d dψ 1 dy
= r (φ + θ ) = r =r Then, y1 (0) = = 0.
dr ds ds R d x x=0, y=0
Now, we have
Therefore, we obtain
x2 x2 2x
dr lim = lim = lim
R=r (8.18) x →0 2y x → 0 2 f (x) x → 0 2 f  (x)
dp y→0

Illustration 8.12 Consider the pedal curve pa = r 2 . [applying L’Hospital’s rule]

dr dr a 1
Here, 2r = a or r = . = lim  [applying L’Hospital’s rule again]
dp dp 2 x → 0 f (x)
Hence, the radius of curvature at any point ( p, r ) is 1 1
a =  =
R= . f (0) y2 (0)
2
Example 8.15 For the pedal curve p2 (a 4 + r 4 ) = a 4 r 2 , find Thus, the radius of curvature at the origin is given by
a  3
the radius of curvature when r = √ . 1 + {y1 (0)}2 2 1 x2
2 R(0, 0) = = = lim
a y2 (0) y2 (0) x → 0 2y
Solution When r = √ , we have from the given equation y→0
2
  Illustration 8.13 Consider the parabola x 2 = 8y. At the
a 4 a2 2a 2 origin, the x-axis is tangent to the parabola.
p2 a 4 + = a4 or p 2 =
4 2 5 Thus, the radius of curvature of the parabola at the origin
√ is
2
Take p = √ a. x2 8y
5 R = lim = lim =4
x → 0 2y x → 0 2y
By differentiation of p 2 a 4 + r 4 = a 4 r 2 w.r.t. p, we have y→0 y→0
 
4 4 2 3 dr dr Case 2 The y-axis is tangent to the curve at the origin.
2 p(a + r ) + p 4r = 2a 4 r
dp dp Then the radius of curvature at the origin is similarly given
dr  4  by
or a r − 2 p2r 3 = p a 4 + r 4
dp y2
R(0, 0) = lim
dr p a4 + r 4 x →0 2x
or = 4 y→0
dp a r − 2 p2r 3
 Applications of Differential Calculus 179

Illustration 8.14 Consider the parabola y 2 = 4ax. At Illustration 8.15 Consider the curve y = e2x − 1. At the
the origin, the y-axis is tangent to the parabola. origin, neither the x-axis nor the y-axis is tangent to the
Thus, the radius of curvature of the parabola at the ori- curve.
gin is We write f (x) = e2x − 1.
y2 4ax Then, f (0) = 1 − 1 = 0, f  (0) = 2e0 = 2, f  (0) = 22 e0 =
R = lim = lim = 2a 4, . . . .
x →0 2x x → 0 2x
y→0 y→0 Expanding by Maclaurin’s theorem, we have
Case 3 Neither the x-axis nor the y-axis is tangent to the x  x 2  x x2
curve at the origin. y = f (0) + f (0) + f (0) + · · · = 2 + 4 + · · ·
1! 2! 1! 2!
By Maclaurin’s theorem, we have
Here, p = 2 and q = 4.
x  x 2  Thus, the radius of curvature at the origin is
y = f (0) + f (0) + f (0) + · · ·
1! 2! 3 √
x  2
x  (1 + p 2 ) 2 5 5
= f (0) + f (0) + · · · [since f (0) = 0] R= =
1! 2! q 4
x x2
= p +q + · · · [where p = f  (0) and q = f  (0)] Polar Form
1! 2! Radius of curvature of a polar curve passing through the
Thus, the radius of curvature at the origin is given by pole can be found as follows.
 3  3  3 We know x = r cos θ and y = r sin θ .
1 + {y1 (0)}2 2 1 + { f  (0)}2 2 1 + p2 2
R(0, 0) = = =
y2 (0) f  (0) q x2 r 2 cos2 θ r
R(0, 0) = lim = lim = lim
x2 x →0 2y r →0 2r sin θ r →0 2θ
where p and q are, respectively, the coefficients of x and y→0 θ →0 θ →0
2
in the Maclaurin’s series for f (x). =
1
lim
dr
2 r →0 d θ
θ →0

Section Review 8.4

1. Find the radius of curvature of the curve y 3 = a 2 x at any 5. Find the radius of curvature of the curve r = 1 − sin θ at
point (x, y) on it. the point (1, 0).
2. Find the radius of curvature at any point of the catenary 6. Find the radius of curvature of the curve r = 2 + sin 3θ
a x x

at the point (2, 0).
y= e a + e− a
2 7. Find the radius of curvature of the curve r = 3 cos θ at
3. Find the radius of curvature of the parabola y 2 = 4 px the point (3, 0).
(a) at any point on the curve and (b) at the vertex. 8. For the curve defined by the parametric equations x =
4. Find the radius curvature of the parabola y 2 = 4x at the 3t, y = t 2 + 2, find the radius of curvature for any t.
vertex.

8.6 Circle of Curvature

Let P be any point on a given curve. Let us construct a the curve at P and P T be the tangent at the point P.
circle such that it touches the curve at point P, it has Suppose the tangent makes angle ψ with the x-axis. Draw
the same curvature as that of the curve, and it lies on the a circle with radius R, touching the tangent P T at P and
same side of the tangent at P, then it is called the circle of on the same side of the tangent as that of the curve. Then,
curvature of the curve at the point P. the centre C(X, Y ) of this circle is the centre of curvature
The centre of the circle of curvature is called the centre (see Fig. 8.6).
of curvature at the point P. From Fig. 8.6, we see that
Let us first find the co-ordinates of the centre of curva- y1 y1 (1 + y12 )
X = x − R sin ψ = x − R  = x −
ture and the circle of curvature in Cartesian co-ordinates. 1 + y12 y2
y1 (1 + y12 )
Y = y + R cos ψ = y + R  =y+
8.6.1 Cartesian Form 1 + y12 y2
Let the equation of a curve be y = f (x) and P (x , y) be Thus, for any point P(x, y) on the curve, the radius of curv-
any point on the curve. Let R be the radius of curvature of ature R and the centre of curvature C (X, Y ) can be found.
180 Engineering Mathematics

8.6.2 Parametric Form

Y
When the equations of a curve are given in parametric form
C(X, Y) x = x(t) , y = y(t), t being the parameter, the co-ordinates
of the centre of curvature are given by
! " ! "
R cos␺ ␺ R y  (x  )2 + (y  )2 x  (x  )2 + (y  )2
X=x− and Y = y +
x  y  − x  y  x  y  − x  y 
y⫽f(x) P(x, y) 2
dx dy d x d2 y
where x  = , y = , x  = 2 , and y  = 2 .
dt dt dt dt
Example 8.17 For the curve defined by the parametric
␺
equations x = 3t, y = t 2 − 6, find the radius of the circle of
O T X
R sin ␺ curvature and the centre of curvature at any point t = 1. Also,
find the equation of the circle of curvature at the point.
Fig. 8.6 Circle of curvature and centre of curvature dx d2x dy d2 y
Solution We have = 3, 2
= 0 and = 2t, = 2.
The equation of the circle of curvature is given by dt dt dt dt 2
dx 2
d x
At t = 1, we have x = 3, y = −5, = 3, = 0, and
(x − X )2 + (y − Y )2 = R 2 dt dt 2
dy 2
d y
Example 8.16 For the parabola y = x 2 , find the  radius
 of = 2, = 2.
1 1 dt dt 2
curvature and the circle of curvature at the point , . Thus, we get
2 4
% &3
dy d2 y x
2
+ y
2 2 3 √
Solution We have = 2x and = 2. (32 + 22 ) 2 13 13
 d x dx2 R= = =
1 1 dy d2 y x  y  − x  y  (3)(2) − (2)(0) 6
At the point , , we have = 1 and = 2.
2 4 dx dx2 The centre of curvature (X, Y ) is given by
Thus 2 2
3  3 x  + y
 (2)(9 + 4) 13 4
1 + y12 2 1 + 12 2 √ X = x − y   =3− =3− =−
R= = = 2 x y − x  y  6 3 3
y2 2 2 2
x  + y (3)(9 + 4) 13 3
The centre of curvature (X, Y ) at the point is given by Y = y + x    
= −5 + = −5 + =
x y −x y 6 2 2
1 1(1 + 12 ) 1 Thus, the circle of curvature is given by
X = − =−
2 2 2 '  (2 ' (2 ' √ (2
1 (1 + 12 ) 1 5 4 3 13 13
Y = + = +1= x− − + y− =
4 2 4 4 3 2 6
 
1 1 16 9
Thus, the circle of curvature at the point , is given by or 36x 2 + 36y 2 + 96x − 108y = −36 × − 36 × + 2207
2 4 9 4
'  (2 ' (2 √ 2 = −64 − 81 + 2207
1 5
x− − + y− = 2
2 4
or 36x 2 + 36y 2 + 96x − 108y = 2062
5 3
or x 2 + y2 + x − y= . or 18x 2 + 18y 2 + 48x − 54y = 1031.
2 16

Section Review 8.5

1. Find the centre of curvature of the curve y 3 = a 2 x at any 5. Find the centre of curvature of the curve r = 2 + sin 3θ
point (x, y) on it. at the point (2, 0). Also, find the equation of the circle of

a x
2. Find the centre ofcurvature at any point of the catenary curvature.
x
y= e a + e− a . 6. Find the centre of curvature and the equation of the
2 circle curvature of the curve r = 3 cos θ at the point
3. Find the co-ordinates of the centre of curvature of the
parabola y 2 = 4 px (a) at any point on the curve and (3, 0).
(b) at the vertex. Also, find the equation of the circle of 7. For the curve defined by the parametric equations
curvature at the vertex. x = 3t, y = t 2 + 2, find the centre of curvature at
4. Find the centre of curvature and the equation of the cir- any point on it.
cle curvature of the curve r = 1 − sin θ at the point (1, 0).
 Applications of Differential Calculus 181

8.7 Evolute d f (x) d 2 f (x)

Since y = f (x), y1 = , and y2 = , the set of
dx dx2
Consider any plane curve 1 . At any point P on 1 , the equations in Eqn (8.19) forms a parametric form of equa-
centre of curvature of the curve can be found. As the point tions of the evolute of the curve y = f (x), x being the
moves along the curve, the centre of curvature may also parameter.
move, and thus, its locus forms another curve 2 . This If x is eliminated from Eqn (8.19), we will get an equa-
curve 2 , which is the locus of the centres of curvature tion involving X and Y , i.e., the Cartesian equation of the
corresponding to the points on the curve, is known as the evolute.
evolute of the curve 1 . The curve 1 is called the involute Thus, to find the evolute of a curve, the following steps
of the second curve 2 . are to be executed:
Suppose, C, C1 , C2 , · · · , C7 are, respectively, the centres
Step 1 Find X and Y from Eqn (8.19).
of curvature at points P, P1 , P2 , · · · , P7 on the curve 1 , as
shown in Fig. 8.7. Then, the curve CC1 C2 C7 is the evolute Step 2 Solve the resulting equations for x and y in terms
2 of the curve P P1 P2 P7 (i.e., 1 ).
of X and Y .

C7 Step 3 Substitute these values of x and y in the given

Evolute equation. This gives a relation between X and Y , which is
the equation of the evolute.
C5
Illustration 8.17 Consider the curve of Example 8.17
and let us find the evolute for this curve.
C4
C1 Any point on the curve is given by x = 3t, y = t 2 − 6.
C dx d2x dy d2 y
We have = 3, 2
= 0 and = 2t, = 2.
P dt dt dt dt 2
Thus, we get
P1 !  2 "3
(x ) + (y  )2 2
3
P2 (32 + 22 ) 2 (9 + 4t 2 )3/2
P3 R= = =
P4 x  y  − x  y  (3)(2) − t 2 (0)(2t) 6
P5 P
6 P7 The centre of curvature C (X, Y ) is given by
Curve (x  )2 + (y  )2 (2t)(32 + 22 t 2 )
X = x − y = 3t −
x  y  − x  y  (3)(2) − (0)(2t)
Fig. 8.7 Evolute of a plane curve
9t + 4t 3 4
= 3t − = − t3
Illustration 8.16 Consider two well known curves (a) a 3 3
parabola and (b) an Ellipse. (x  )2 + (y  )2 (3)(32 + 22 t 2 )
Y = y + x = t 2
− 6 +
The evolutes of these two curves are shown in Fig. 8.8. x  y  − x  y  (3)(2) − (0)(2t)
9 + 4t 2 3
= t2 − 6 + = 3t 2 −
2 2
Evolute Eliminating the parameter t from X and Y , we have
⎡ ⎤
Ellipse 3 3 ' (
Y+ 3 2
⎢ 2 ⎥
⎣ ⎦ = − X or 2(2Y + 3)3 = 243X 2
3 4
Evolute Thus, the equation of the evolute of the curve defined by
Parabola x = 3t, y = t 2 − 6 is 2(2y + 3)3 = 243x 2 .

 Notes It can be proved that

(a) (b)
1. At any point on a curve, the radius of curvature is equal
Fig. 8.8 Evolutes of (a) a parabola and (b) an ellipse to the radius of the circle of curvature;
2. The length of arc of the evolute corresponding to points
If y = f (x) be the equation of the curve 1 in Carte- on the curve between two points P1 and P2 is the differ-
sian co-ordinates, then the co-ordinates (X, Y ) of centre of ence between radii of curvature R1 and R2 of the curves
curvature at point P(x, y) on the curve are given by at the points P1 and P2 ; and
y1 (1 + y12 ) (1 + y12 ) 3. Normal at any point to the curve is tangent to the
X=x− and Y = y + (8.19)
y2 y2 evolute of the curve.
182 Engineering Mathematics

Section Review 8.6

x2 y2
1. Find the evolute of the hyperbola 2 − 2 = 1. (a) x = a(cos t + t sin t), y = a(sin t − t cos t)
a b (b) x = a cos3 t, y = a sin3 t
2. Find the parametric equations of the evolutes of the fo-
(c) x = 2t, y = t 2 − 2
llowing parametric curves.

8.8 Envelope
Y
Let f (x, y, α) = 0 be an equation, where α is a parameter.
Corresponding to a particular value of the parameter, there
Envelope
exists a plane curve, i.e., different values of α yield different
B 2 2
curves. Thus, f (x, y, α) = 0 represents a family of curves x 3 ⫹y 3 ⫽1
in the x y-plane, where α is a parameter and therefore, it is
called a one-parameter family of curves. Similarly, an equa- d P
tion of the form f (x, y, α, β) = 0, where α and β are two
␪
parameters, represents a two-parameter family of curves.
O A X
 Note If α and β are connected by a relation, i.e.,
one is dependent on the other, the two-parameter family of Fig. 8.9 Envelope of a slipping line
curves reduces to a one-parameter family of curves.
Envelope For a given family of curves, an envelope, if it and
exists, is defined to be a curve which touches every member ∂ x sin θ y cos θ
f (x, y, θ) = − =0
of the family of curves and at each of its points, it is touched ∂θ cos θ
2 sin2 θ
by some member of the family of curves. Solving these equations, we get x = d cos3 θ, y = d sin3 θ .
The following are the steps, without proof, to find the These are the parametric equations of an asteroid.
envelope of a one-parameter family of curves. Eliminating the parameter θ, we can express the equa-
tion of the envelope in Cartesian co-ordinates as
Step 1 Let f (x, y, α) = 0 be the equation, where α is the 2 2
parameter. Construct another equation by differentiating it x3 + y3 = 1
∂
w.r.t. α, i.e., f (x, y, α) = 0. Example 8.18 Find the envelope of co-axial ellipses, the
∂α
Step 2 Eliminate the parameter α from these two equa- sum of whose major and minor axes is a constant, say d.
∂
tions, i.e., f (x, y, α) = 0 and f (x, y, α) = 0. Solution Suppose the major and minor axes of an ellipse
∂α x2 y2
are a and b, respectively, then its equation is 2 + 2 = 1.
Step 3 The resulting equation is the Cartesian equation a b
of the envelope of the family of curves. It is given that a + b = d.
x2 y2
The equation 2 + 2 = 1 represents a one-parameter family
 Note If x and y are solved in terms of the parameter a b
α, the parametric equations of the envelope are obtained. of curves, as the two parameters a and b are bounded by the
relation a + b = d. Let t be the parameter.
Illustration 8.18 Consider a segment AB of fixed length Then, we have
d, slipping on x-axis and y-axis as shown in Fig. 8.9. Let
da db db da
us find the envelope of this variable segment. a+b =d ⇒ + =0⇒ =−
dt dt dt dt
The x co-ordinate of A is in [0, d]. When AB makes an
To obtain the parametric equations of the envelope, it is re-
angle θ with AO, the co-ordinates of A are (d cos θ, 0) and
quired to eliminate t (i.e., a and b) from
those of B are (0, d sin θ ).
Thus, the equation of the line AB is x2 y2
f (x, y, t) = + =1 (8.20)
x y a2 b2
f (x, y, θ) = + =d
cos θ sin θ Now
To obtain the parametric equations of the envelope, it is ∂f −2x 2 da −2y 2 db
required to solve for x and y in terms of θ in the following = + =0
∂t a 3 dt b3 dt
system of equations:
x y x2 y2
f (x, y, θ) = + =d or − =0 (8.21)
cos θ sin θ a3 b3
 Applications of Differential Calculus 183

From Eqn (8.20) and Eqn (8.21), we get dY dy

But is the slope of the tangent to the evolute and
a3 b3 dX dx
x2 = and y 2 = is the slope of the tangent to the curve.
d d
1 1 Therefore, a normal to the given curve is a tangent to
or a = dx2 3
and b = dy 2 3
its evolute.
a and b are eliminated by adding these equations Thus, an evolute may be alternately defined as the en-
1 2 1 2 2 2 2 velope of the normals to the curve.
d = d 3 x 3 + d 3 y 3 or x 3 + y 3 = d 3
Illustration 8.19 Consider the parabola y = ax 2 , where
The above equation represents the Cartesian equation of an
a is a positive constant. A variable point of this parabola
asteroid.
is P (t, at 2 ).
The tangent line in P has slope 2at. The normal has
8.8.1 Evolute as Envelope of Normals 1
slope − .
As already discussed, the co-ordinates of the centre of cur- 2at
vature C(X, Y ) of a curve y = f (x) are given by The set of all normals can be written as
1
y − at 2 = − (x − t)
X = x − R sin ψ, Y = y + R cos ψ 2at
or 2at y − 2a 2 t 3 + x − t = 0 or (2ay + 1)t − 2a 2 t 3 + x = 0
Choose the arc-length of the curve s as an independent
To obtain the parametric equations of the envelope to the
variable, then x, y, R, X, Y are functions of s.
family of normals, we calculate x and y from the system of
Differentiating w.r.t. s, we get
equations
dX dx dψ dR
= − R cos ψ − sin ψ (8.22) (2ay + 1)t − 2a 2 t 3 + x = 0
ds ds ds ds
(2ay + 1) −6a 2 t 2 = 0
dY dy dψ dR
= − R sin ψ + cos ψ (8.23) We find
ds ds ds ds 1
From Note 2 of Section 8.3.2, we have x = −4a 2 t 3 , y = 3at 2 +
2a
dx dy dψ 1 To obtain the Cartesian equation of the envelope, eliminate
= cos ψ, = sin ψ, = the parameter t and we find the semi-cubic parabola
ds ds ds R  
1 3
Substituting in Eqn (8.22) and Eqn (8.23), we get 16a y − = 27x 2
2a
dX 1 dR dR
= cos ψ − R cos ψ − sin ψ = − sin ψ (8.24) This is the equation of evolute of the parabola y = ax 2 .
ds R ds ds
 Notes
dY 1 dR dR 1. The envelope passes through any ordinary cusp-point
= sin ψ − R sin ψ + cos ψ = cos ψ (8.25)
ds R ds ds of the original curve.
Dividing Eqn (8.25) by Eqn (8.24) gives 2. Points of inflexion on the original curve correspond to
dY 1 1 points at infinity on the evolute.
= − cot ψ = − =− . 3. Points of maximum or minimum curvature of the orig-
dX tan ψ dy
dx inal curve correspond to cusps of the evolute.

Section Review 8.7

1. Determine the envelope of x sin t − y cos t = a, where t is 5. Find the envelope of the family of parabolas
a parameter. y 2 = 4a(x − a), where a is a parameter.
2. Find the envelope of the family of straight lines 6. Show that the envelope of the family of straight lines
y = mx + am 2 , where m is a parameter. x sec2 θ + ycosec2 θ = a, where θ is a parameter is given
3. Find the envelope of the family of straight lines by (x − y)2 − 2a(x + y) + a 2 = 0.
 7. Show that the envelope of the family of straight lines
y = mx + 1 + m2
x cos θ + y sin θ = a cos θ sin θ, where θ a parameter, is
2 2 2
where m is a parameter. given by x 3 + y 3 = a 3 .
4. Find the envelope of the family of straight lines 8. Find the envelope of the family of circles described on
 x2 y2
y = mx + a 2 m 2 + b2 double ordinates of the ellipse 2 + 2 = 1 as diameter,
a b
where a and b are constants.
where m is a parameter.
184 Engineering Mathematics

8.9 Curve Tracing Asymptotes

An asymptote to a plane curve is a straight line which
It is sometimes helpful to know an approximate shape of a
touches the curve at infinity.
curve, particularly in applications that involve finding its
arc-length, area or volume enclosed by it, etc. Characteris- (i) Asymptotes parallel to the x-axis These are obtained
tics like symmetry, asymptotes, intercepts, etc. of a curve by equating to zero the coefficient of the highest powers
are important for finding its approximate shape. Let us of x in the equation f (x, y) = 0. If the coefficient of
consider some of these important characteristics for curves the highest powers of x in the equation is a constant,
in Cartesian co-ordinates, as well as polar co-ordinates and the curve does not have any asymptote parallel to the
for parametric curves. x-axis.
(ii) Asymptotes parallel to the y-axis These are obtained
8.9.1 Cartesian Equation by equating to zero the coefficient of the highest powers
Consider a function f (x, y) of the form f (x, y) = y n + of y in the equation f (x, y) = 0. If the coefficient of
u 1 (x)y n−1 + u 2 (x)yn−2 + · · · + u n (x), where u r (x) is a poly- the highest powers of y in the equation is a constant,
nomial in x of maximum degree r (r = 1, 2, . . . , n). Then the curve does not have any asymptote parallel to the
the equation f (x, y) = 0 represents a plane curve. The y-axis.
curve is called a plane algebraic curve of degree n. (iii) Oblique asymptotes An asymptote which is not par-
allel to any of the co-ordinate axes is known as an
 Note Some properties of a curve are better under- oblique asymptote.
stood when its equation is expressed in the above form. Oblique asymptotes can be obtained as follows:
Given below are some of the characteristics of a curve Suppose the equation of the curve can be explicitly
expressed in the Carterian form. written in the form y = f (x). If
y
m = lim and c = lim (y − mx)
x→∞ x x→∞
Symmetry
Then y = mx + c is an oblique asymptote to the curve.
(i) A plane algebraic curve f (x, y) = 0 is symmetric When the equation of the curve is given in implicit
about the x-axis if it contains only even powers of y so form f (x, y) = 0, rewrite the equation as
that it remains unchanged if y is replaced by −y, i.e.,
f (x, −y) = f (x, y). Similarly, a plane algebraic curve f (x, y) = f n (x, y) + f n−1 (x, y) + g(x, y) = 0
f (x, y) = 0 is symmetric about the y-axis if it contains where f n (x, y) consists of all the terms of (the highest)
only even powers of x so that it remains unchanged if degree n, f n−1 (x, y) consists of all the terms of (the
x is replaced by −x, i.e., f (−x, y) = f (x, y). next highest) degree (n − 1), and g(x, y) consists of
(ii) A plane algebraic curve f (x, y) = 0 is symmetric other lower degree terms.
about both the x-axis and y-axis if it contains only Let p(m) = f n (1, m) and q(m) = f n−1 (1, m).
even powers of y and x so that it remains unchanged Then y = mx + c is an oblique asymptote, where m is
if x is replaced by −x and y is replaced by −y simul- a solution of p (m) = 0 and c is a solution of cp  (m) +
taneously, i.e., f (−x, −y) = f (x, y). q (m) = 0.
(iii) A plane algebraic curve f (x, y) = 0 is symmetric Illustration 8.21 The axes x = 0 and y = 0 are both
about the origin if f (−x, −y) = f (x, y). asymptotes to the curve defined by 4x y = c2 .
(iv) A plane algebraic curve f (x, y) = 0 is symmetric x2 y2
The curve defined by 2 − 2 = 1 has the oblique asymp-
about the line y = x if the equation remains unaltered a b
b
when x and y are interchanged, i.e., f (y, x) = f (x, y). totes y = ± x.
a
(v) A plane algebraic curve f (x, y) = 0 is symmetric
about the line y = −x if the equation remains unal- Intercepts
tered when −y is written in place of x and −x in place The x intercepts of the curve, i.e., the points of intersection
of y, i.e., f (−y, − x) = f (x, y). of the curve with the x-axis are obtained by putting y = 0
in the equation and then solving for x. In a similar way, y
Illustration 8.20 The curve defined by y 2 = 4ax is intercepts are found.
symmetric about the x-axis and the curve defined by
Illustration 8.22 For the curve defined by
x2 y2
+ 2 =1 x2 y2
a 2 b + =1
a2 b2
is symmetric about both x-axis and y-axis. the x-intercepts are x=±a and the y-intercepts are y=±b.
 Applications of Differential Calculus 185

dy d2 y
Curve Passing through Origin
 Note For a stationary point (where = 0), if
Whether the curve passes through the origin or not can dx dx2
< 0, the curve has a maxima at the point. Similarly, for
be easily found by inspecting the expression f (x, y). If d2 y
f (x, y) does not contain any constant term, then the curve a stationary point, if > 0, the curve has a minima at
dx2
passes through the origin. that point.
If a curve passes though the origin, it is important to find Illustration 8.25 Consider again the curve y = x 3 −12x.
the tangent to the curve at the origin. This can be obtained dy
We have seen that = 3x 2 − 12 = 0 for x = −2 and
by equating the group of lowest degree terms to zero. dx
x = 2, and hence, x = −2 and x = 2 are stationary points
Illustration 8.23 The curve y 2 = 4ax passes through of the curve.
x2 y2 d2 y
the origin, whereas the curve 2 + 2 = 1 does not. Now, = 6x.
a b dx2 2
d y
Sign of First and Second Derivatives For x > 0, is positive and the curve is convex.
dx2
For curve tracing, it is important to know the maxima and d2 y
minima of the curves, as well as the increasing and de- For x < −2, is negative, and hence, the curve is
dx2
creasing properties of it in different sub-intervals. For this concave.
dy d2 y
purpose, it is required to check the first and second deriva- Thus, at x = 2, = 0 and is positive, and hence,
dx dx2
tives of one variable w.r.t. the second one in the following
the curve has a minimum at x = 2. Similarly, the curve has
manner.
a maximum at x = −2.
dy
Sign of
dx
dy Extent of the Curve
From the equation of the curve, find . In an interval
dx For a given curve f (x, y) = 0, solve for y in terms of x
dy
[a, b], if > 0, the curve is increasing in the interval, and find the values of x for which y is defined. This gives
dx
dy the horizontal extent of the curve. Similarly, solve for x in
and if < 0, the curve is decreasing in the interval.
dx terms of y and find the values of y for which x is defined.
dy
If = 0 for x = x0 , then at x = x0 , the curve is sta- This gives the vertical extent of the curve.
dx
tionary.
Multiple Points
Illustration 8.24 Consider the curve y = x 3 − 12x.
dy If multiple branches of a curve pass through a point, it is
We have = 3x 2 − 12. called a multiple point. A point through which r branches
dx
dy of a curve pass is known as a multiple point of order r and
Now, = 0 gives x 2 − 4 = 0, i.e., x = 2 or x = −2.
dx the curve has r tangents at that point.
dy
For x < −2, > 0, and hence, the curve is increasing for A multiple point of order 2 is known as a double point.
dx
x < −2. A double point is further classified as a node, a cusp,
dy or an isolated point according to whether the two tangents
For x > 2, > 0, and hence, the curve is increasing
dx at that point are real and distinct, real and coincident, or
for x > 2.
dy imaginary.
For −2 < x < 2, < 0, and hence, the curve is
dx Multiple points of a curve f (x, y) = 0 can be obtained
decreasing for −2 < x < 2. by solving the equations f (x, y) = 0, f x (x, y) = 0, and
Thus, the curve has maximum at x = −2 and has a mini- f y (x, y) = 0.
mum at x = 2. At a double point, the slopes of the tangents are the
d2 y roots of the following equation in m: f yy (x, y) m 2 +2 f x y (x, y)
Sign of m + f x x (x, y) = 0.
d x2
d2 y Now, the values of m (roots of the equation) are real
In an interval [a, b] for a given curve, if > 0, the curve
dx2 and distinct, real and equal, or imaginary depending on
d2 y  2
is convex in the interval and if 2 < 0, the curve is concave whether f x y (x, y) − f yy (x, y) f x x (x, y) > 0, = 0, or < 0,
dx respectively.
in the interval.
2
d y Thus, a double point (x, y) is, respectively, a node, a
If = 0 for x = x0 , then at x = x0 , the curve changes
dx2 cusp, or an isolated point if
from being a convex one to a concave one, or vice versa.  2
Such a point is known as an inflection point. f x y (x, y) − f yy (x, y) f x x (x, y) > 0, = 0, or < 0.
186 Engineering Mathematics

Illustration 8.26 Consider the equation f (x, y) = x 3 − y 2 (d) Origin does not lie on the curve.
= 0. It describes a plane curve, as shown in Fig. 8.10, known (e) dy
=
d 3
(x − 4x 2 + x + 6) = 3x 2 − 8x + 1 = 0
as the semi-cubical parabola. dx dx √
8 ± 64 − 12
for x = = 2.535, 0.131.
Y 6

x3⫺y2 ⫽0 Thus, x = 0.131, 2.535 are stationary points.

d2 y
(f) = 6x − 8 = 0 for x = 0.131, 2.535 . Thus, the curve
(0, 0) dx2
has no inflection point.
O X d2 y
At x = 0.131, < 0, and hence, y is maximum at
dx2
that point.
Maximum value is [y]x=0.131 = 6.06.
Fig. 8.10 Multiple point: a cusp d2 y
At x = 2.535, > 0, and hence, y is minimum at
The origin (0, 0) lies on the curve. Let us check this dx2
that point.
point of the curve. Minimum value is [y]x=2.535 = −0.88.
f x (x, y) = 3x 2 , f y (x, y) = −2y, f x x (x, y) = 6x, f x y (x, y) (g) Extent of the curve is (− ∞, ∞).
= 0, and f yy (x, y) = −2. dy
(h) Sign of y and in the various integrals are given in the
At the origin (0, 0), f x x (x, y) = 0, f x y (x, y) = 0, and following table:
dx
f yy (x, y) = −2.
 2
Thus, f x y (x, y) − f yy (x, y) f x x (x, y) = 02 −0(−2) = 0. Sign Increasing/
dy
Hence, the point (0, 0) is a cusp. Interval of x of y Quadrant Sign of
dx
decreasing
Now, consider another equation
−∞ < x < −1 −ve Third +ve Increasing
f (x, y) = x 3 − y 2 − 3x + 2 = 0. −1 < x <0 +ve Second +ve Increasing
The curve is shown in Fig. 8.11. 0<x < 0.131 +ve First +ve Increasing
0.131 < x <2 +ve First −ve Decreasing
Y 2<x < 2.535 −ve Fourth −ve Decreasing
2.535 < x <3 −ve Fourth +ve Increasing
3<x <∞ +ve First +ve Increasing
(1, 0)
The curve passes through the points (−1, 0), (0, 6), (0.131,
O X
6.06), (2, 0), (2.535, −0.88), and (3, 0).
From the above information about the curve, the curve can
be traced as shown in Fig. 8.12.
Fig. 8.11 Multiple point: a node

We see that the point (1, 0) lies on the curve. Let us y (0.131, 6.06)
check this point of the curve. 6
f x (x, y) = 3x 2 − 3, f y (x, y) = −2y, f x x (x, y) = 6x,
f x y (x, y) = 0, and f yy (x, y) = −2. 5
At the point (1, 0), f x x (x, y) = 6, f x y (x, y) = 0, and
f yy (x, y) = −2. 4
Thus
 2 3
f x y (x, y) − f yy (x, y) f x x (x, y) = 02 − (6)(−2) = 12 > 0
Hence, the point (1, 0) is a node on this curve. 2
Example 8.19 Trace the curve y = (x + 1)(x − 2)(x − 3).
1
Solution The curve y = (x + 1)(x − 2)(x − 3) has the fo-
llowing characteristics:
(a) The curve has no symmetry. O 1 2 3 4 x
(b) It has no asymptotes.
(c) The x-intercepts of the curve are −1, 2, 3, and the
y-intercept is 6. Fig. 8.12 Curve for y = (x + 1)(x − 2)(x − 3)
 Applications of Differential Calculus 187

2
Example 8.20 Trace the curve x 3 + y 3 = 3ax y. Since f x y − f x x f yy = 9a 2 − 36x y > 0 for x = 0,
y = 0, the point (0, 0) is a node of the curve.
Solution The equation of the curve is f (x, y) ≡ x 3 + y 3 − dy
3ax y = 0. It has the following characteristics: The values of y and for different values of x are not
dx
easy to compute. However, from the above information
(a) As f (y, x) = f (x, y), the curve is symmetric about the
about the curve, the curve can be traced as shown in Fig.
line y = x.
8.13.
(b) (i) The coefficients of highest power of x and also of y
are constants, and hence, the curve does not have any
asymptote parallel to the co-ordinate axes. Y
(ii) Now, let us find oblique asymptotes, if any.
Rewrite the equation of the curve as  3a2 , 3a2 
f (x, y) = (x 3 + y 3 ) − 3ax y = 0
y⫽x
If y = mx + c is an oblique asymptote to the curve,
then m is the real value solution of x3⫹y3 ⫽ 3axy

(x 3 + y 3 ) x=1,y=m
= 0, i.e., 1+m 3 = 0, i.e., m = −1
O (0, 0) X
and c is obtained from
d a x⫹y⫽a
(1 + m 3 )c − 3am = 0, i.e., c = = −a .
dm m
Thus, the curve has an oblique asymptote y = −x − a,
i.e., y + x + a = 0.
(c) The curve intersects the co-ordinate axes only at the origin
(0, 0). The points of intersections of the curve with the line
of symmetry y = x are obtained from x 3 + x 3 = 3ax x,
3a
which gives x = 0, . Thus, the points are (0, 0) and
  2 Fig. 8.13 Curve for x 3 + y 3 = 3ax y
3a 3a
, .
2 2 1
(d) The origin lies on the curve. Example 8.21 Sketch the curve for the equation y = x + .
x
(e) Equating the coefficient of lowest degree term (i.e., 3ax y)
Solution Rewrite the equation of the curve as f (x, y) =
to zero, we see that at the origin, both the axes x = 0 and
x y − x 2 − 1 = 0.
y = 0 are tangents to the curve.
The curve shows the following characteristics:
Also, differentiating the equation of the curve, we have
dy dy (a) The curve is not symmetric about either the x-axis or
3x 2 + 3y = 3ay + 3ax y-axis. However, it is symmetric about the origin, as
dx dx
f (−x,−y) = (−x)(−y) − (−x)2 − 1 = x y − x 2 − 1 = f (x, y)
dy ay − x 2
or = 2 (b) Equating the coefficient of highest power of y to zero, we
dx y − ax
see that the curve has a vertical asymptote x = 0.
  1
3a 3a dy From y = x + , we see that y → −∞ as x → 0 from
At , = −1 x
2 2 dx negative side of x-axis and y → +∞ as x → 0 from
So, the tangent at the point is parallel to the asymptote positive side of x-axis.
y + x + a = 0. Since the coefficient of highest power of x (i.e., coeffi-
dy ay − x 2 x2 x6 x2 cient of x 2 ) is constant, the curve has no horizontal asymp-
(f) = 2 = 0 when y = , i.e., x 3 + 3 = 3ax
dx y − ax a a a tote.
√
6 3 3
i.e., x = 2a x , i.e., x = 3 2a. Now, let us search for oblique asymptotes.
(g) On the curve x 3 + y 3 − 3ax y = 0, x and y cannot be both If y = mx + c is an asymptote, then
 
negative, because when both x and y are negative, each y 1
m = lim = lim 1 + 2 =1
of the terms x 3 , y 3 , and −3ax y are negative, and hence, x→∞ x x→∞ x
x 3 + y 3 − 3ax y cannot be zero. and
(h) Since f (x, y) = x 3 + y 3 −3ax y = 0, f x (x, y) = 3x 2 −3ay = 1
lim (y − mx) = lim=0
0, and f y = 3y 2 −3ax = 0 has only one solution x = 0, y = x→∞ xx→∞
0, the curve has one multiple point (x = 0, y = 0). Thus, the curve has an oblique asymptote y = x.
188 Engineering Mathematics

1 Example 8.22 Sketch the curve for the equation

(c) The equation of the curve y = x + is not defined for
x
x = 0 and also y cannot be zero for any value of x.
x 2 + 3x − 1
Thus, the curve does not intersect any of the axes. y= .
(d) From (c), it can be observed that the origin does not lie x −2
on the curve.
dy 1 x2 − 1
(e) =1− 2 = = 0 for x = ±1. Solution Rewrite the equation of the curve as
dx x x2
Thus, x = −1, 1 are stationary points. 9
y = x +5+ .
d2 y 2 x −2
(f) 2
= 3 =  0 for x = −1 or x = 1. Thus, the curve has
dx x The following observations can be made about the curve:
no inflection point.
(a) The curve is symmetric neither about the x-axis nor
d2 y
For x = −1, < 0, and hence, y is maximum at about the y-axis. It is also not symmetric about the
dx2
x = −1 and [y]x=−1 = −1 + −1 1
= −2. origin.
(b) Equating the coefficient of highest power of y to zero, we
d2 y see that the curve has a vertical asymptote: x = 2.
For x = 1, > 0, and hence, y is minimum at x = 1
dx2 Now let us search for existence of oblique asymptotes.
and [y]x=1 = 2.
If y = mx + c is an asymptote, then
(g) The curve is defined for x ∈ R − {0}. Also, signs of x and
y must be same, so the curve lies in the first and the third  
y 5 9
quadrants only. m = lim = lim 1 + + =1
x→∞ x x→∞ x x(x − 2)
(h) Since f (x, y) = x y − x 2 − 1 = 0, f x (x, y) = y − 2x = 0,
and f y = x = 0 has no solution, the curve has no multiple and
point.  
dy 9
(i) Signs of y and may be noted as follows. c = lim (y − mx) = lim 5+ =5
dx x→∞ x→∞ x −2
Sign Increasing/
dy Thus, y = x + 5 is an oblique asymptote to the curve.
Interval of x of y Quadrant Sign of decreasing
dx 9
(c) The equation of the curve y = x + 5 + is not defined
−∞ < x < −1 −ve Third +ve Increasing x −2
−1 < x <0 −ve Third −ve Decreasing for x = 2. √
−3 ± 9 − 4
0<x <1 +ve First −ve Decreasing y = 0 when x 2 + 3x − 1 = 0, i.e., x = ,
2
1<x <∞ +ve First +ve Increasing i.e., x = 0.303, −3.3.
Thus, the curve intersects the x-axis at the points
Using all the above information, the curve can be traced as (−3.3, 0) and (0.303, 0).
shown in Figure 8.14. 1
When x = 0, we have y = . Hence, the curve inter-
  2
1
sects the y-axis at 0, .
5 2

4 (d) The origin does not lie on the curve.

dy 9 (x + 1)(x − 5)
3 (e) =1− = = 0 for x = −1, 5.
dx (x − 2)2 (x − 2)2
2
Thus, x = −1, 5 are stationary points on the curve.
1 d2 y 18
(f) = =
 0 for x = −1 or x = 5. Hence, the curve
dx2 (x − 2)3
–4 –3 –2 –1 0 1 2 3 4 5 has no inflection point.
–1
d2 y 2
–2 For x = −1, 2
= − < 0, and hence, y is maximum
dx 3
–3 9
at x = −1 and y = [y]x=−1 = −1 + 5 + = 1.
−3
–4 2
d y 2
For x = 5, = > 0, and hence, y is minimum at
–5 dx2 3
9
x = 5 and y = [y]x=5 = 5 + 5 + = 13.
1 3
Fig. 8.14 Curve for y = x +
x (g) Region of x for which the curve is defined is R − {2}.
 Applications of Differential Calculus 189

(h) Since f (x, y) ≡ y(x − 2) − x 2 − 3x + 1 = 0, f x (x, y) = line passing through the pole and perpendicular to the ini-
y − 2x − 3 = 0, and f y = x − 2 = 0 has no solution, the tial line if
curve has no multiple point.  π   π 
F r, − θ = F r, + θ or F(r, θ) = F(r, π − θ )
dy 2 2
(i) Signs of y and are as follows.
dx A curve F(r, θ) = 0 is symmetric about the pole if F(−r,
θ) = F(r, θ ), i.e., if the equation contains only even powers
Sign
dy
Increasing/ of r .
Interval of x of y Quadrant Sign of decreasing
dx Illustration 8.27 The curve r = 2 cos 3θ is symmetric
−∞ < x < −3.3 −ve Third +ve Increasing about the initial line, whereas the curve r = sin 2θ is sym-
−3.3 < x < −1.0 +ve Second +ve Increasing metric about the line passing through the pole and perpen-
−1.0 < x < 0.0 +ve Second −ve Decreasing dicular to the initial line.
0.0 < x < 0.303 +ve First −ve Decreasing Now consider the curve r 2 = 3 cos2 θ. It is symmetric
0.303 < x < 2 −ve Fourth −ve Decreasing about the pole.
2<x <5 +ve First −ve Decreasing
5<x <∞ +ve First +ve Increasing Pole on the curve
If F(0, θ0 ) = 0 for some θ0 , then the pole lies on the curve.
From all the above information, the curve can be traced as
The tangent to the curve at pole is the line θ = θ0 .
shown in Fig. 8.15.
Illustration 8.28 Consider the polar curve r = 2 cos 3θ.
π
The pole lies on r = 2 cos 3θ, as r = 0 for θ = . Thus,
π 2
50 θ= is a tangent to the curve at the pole.
2
40 Similarly, the pole lies on r = sin 2θ , as r = 0 for θ = 0.
Thus, θ = 0 is tangent to the curve at the pole.
30 Now, consider the curve r 2 = a 2 . The pole does not lie
on it.
20

10 Points of Intersection
Intersection points of the curve with the initial line are
obtained by putting θ = 0 in the equation of the curve and
–4 –2 0 2 4 5 6 8
solving for r .
–10
Intersection points of the curve with any inclined line
–20 θ = θ1 are obtained by putting θ = θ1 in the equation of
the curve and solving for r .
–30
Illustration 8.29 Consider the polar curve r = 2 cos 3θ.
Putting θ = 0, we have r = 2. So, the curve cuts the initial
x 2 + 3x − 1
Fig. 8.15 Curve for y = line at the distance 2.
x −2 π
Consider an inclined line θ = .
π 4
Putting θ = in the equation, we have
8.9.2 Polar Equation 4 π  √
3π
The general equation of a curve in polar co-ordinates (r, θ ) r = 2 cos = −2 cos =− 2
4 4  √ π
is F(r, θ) = 0, which may be expressed in explicit form as So, the curve intersects the inclined line at − 2, .
r = f (θ ) or θ = g(r ). Let us define below a few character- 4
istics of the curve as expressed in polar form. Asymptotes
The curve has an asymptote if there exists a θ = θ1 such
Symmetry that lim r = ∞. If the equation of the curve can be ex-
θ→θ1
A curve F(r, θ) = 0 is symmetric about the initial line, 1
i.e., θ = 0 if F(r, −θ ) = F(r, θ). pressed in the explicit form r = , θ1 can be found solv-
f (θ )
The curve is symmetric about a line θ = α passing ing the equation f (θ ) = 0 and the equation of the asymp-
through the pole if F(r, α − θ ) = F(r, α + θ ) or F(r, θ) = tote corresponding to θ = θ1 is given by r sin(θ − θ1 ) =
π
F(r, 2α−θ ). Thus, the curve is symmetric about θ = ,the f  (θ1 ).
2
190 Engineering Mathematics

Illustration 8.30 Consider the curve with the equation region bounded by two concentric circles is generally called
3 an annulus.
r= .
1 + 2 cos θ The extent of θ can be found from the condition that
1 + 2 cos θ
Writing f (θ ) = , f (θ ) = 0 implies either θ = for these values of θ, the values of r are real.
3
If r = f (θ ) is a periodic function of θ , to find the extent
2π 2π
= θ0 , say, or θ = − = θ1 , say. of r , it is sufficient to consider one period of the function
3 3
Thus, θ = θ0 and θ = θ1 give the directions of the two only.
3
asymptotes to the curve r = .
1 + 2 cos θ   Illustration 8.31 Consider the curve r = 3 + 2 cos θ.
 −2 sin θ   2π 1
Now, f (θ ) = . Thus, f (θ0 ) = f = −√ The minimum and maximum values of cos θ are −1 and 1,
 3  3 3
2π 1 respectively, and hence, values of r lie between 1 and 5.
and f  (θ0 ) = f  − =√ . The curve lies within the annulus region bounded by the
3 3
Therefore, the asymptotes are given by circles of radii 1 and 5.
   
2π 2π 1
r sin θ − = f = −√
3 3 3 Derivative
    If
dr
> 0, then r increases as θ increases and if
dr
< 0,
2π 2π 1
and r sin θ + = f − =√ dθ dθ
3 3 3 then r decreases as θ increases.

Alternatively, we can express them as Loop

√ 1 √ 1 A curve of the form r = a sin nθ or r = a cos nθ , where n is
3r = −   and 3r =   an integer, has n or 2n equal loops according to whether n
2π 2π
sin θ − sin θ + is odd or even, respectively. Each quadrant can be divided
3 3
into n equal parts and in each quadrant, r can be plotted
Fig. 8.16 shows the curve and its two asymptotes. for θ .
If a curve intersects the initial line at A and B and the
curve is symmetric about the initial line, the curve has a
loop between A and B.
3
3 Example 8.23 Sketch the curve r 2 = a 2 cos 2θ, where a
is a positive number. [This curve is known as lemniscate of
Bernoulli.]

r sin 1 Solution The equation of the curve can be rewritten as

3 3 F(r, θ) = r 2 − a 2 cos 2θ = 0.
The following observation can be made about the curve:
2 4 (a) The curve is symmetric about the initial line, as F(r, −θ)
= F(r, θ ).
r sin 1
3 Now, we have
3
 π   π 
F r, − θ = r 2 − a 2 cos 2 − 2θ
2 2
3 3
= r 2 − a 2 cos(π − 2θ )
= r 2 − a 2 cos {2π − (π − 2θ )}
3  π   π 
3 = r 2 − a 2 cos 2 + 2θ = F r, +θ
Fig. 8.16 Asymptotes to the curve r = 2 2
1 + 2 cos θ
π
So, the curve is also symmetric about the line θ = .
2
Extent of the curve Again, the curve is symmetric about the pole, as F(−r, θ )
If the minimum and maximum values of r are a and b, = F(r, θ ).
π
respectively, then the curve lies within the region bounded (b) r = 0 for cos 2θ = 0, i.e., θ = ± . Hence, the pole lies on
4
by the circles of radii a and b, both centred at the pole. A the curve.
 Applications of Differential Calculus 191

(c) The curve has no asymptote, as r is finite for every value The values of r w.r.t θ may be noted as follows.
of θ .
(d) Putting θ = 0 in the equation of the curve and solving for Value Value
r , we find that the points of intersection of the curve with of θ cos θ r=a(1+cos θ ) of θ cos θ r=a(1+cos θ)
the initial line are (−a, 0) and (a, 0).
π π π π 7π
(e) cos 2θ ≥ 0 for − ≤ 2θ ≤ , i.e., for − ≤ θ ≤ and 0 1.000 2.000 a
6
–0.866 0.134 a
2 2 4 4
π 3π π 3π π 5π
cos 2θ < 0 for < 2θ < , i.e., for <θ < . 0.866 1.866 a –0.707 0.293 a
2 2 4 4 6 4
The equation of the curve is r 2 = a 2 cos 2θ , and hence, π 4π
π 0.707 1.707 a –0.500 0.500 a
the curve does not exist between the lines θ = and 4 3
4
3π π 3π
θ= . 0.500 1.500 a 0.000 1.000 a
4 3 2
Using the above information about the curve, the curve π 5π
can be drawn as shown in Fig. 8.17. 0.000 1.000 a 0.500 1.500 a
2 3
2π 7π
–0.500 0.500 a 0.707 1.707 a
3 4
3π 11π
–0.707 0.293 a 0.866 1.866 a
␪⫽␲Ⲑ2 4 6
5π
␪⫽␲Ⲑ4 –0.866 0.134 a 2π 1.000 2.000 a
6
π –1.000 0.000

Using the above information about the curve, the curve can
r 2 ⫽a2 cos2 ␪ be drawn as shown in Fig. 8.18.

(⫺a, 0) (a, 0) X
1.50a, ␲3 
O
1.707a, ␲4 
a, ␲2 
1.866a, ␲6 
0.50a, 2␲
3 
␪⫽⫺␲Ⲑ4
5␲
␪⫽
6
Fig. 8.17 Curve for r 2 = a 2 cos 2θ
O a 2a X
7␲
␪⫽
6
Example 8.24 Sketch the curve r = a(1 + cos θ ). [This
4␲
curve is known as cardioid.] ␪⫽
3 11␲
␪⫽
6
Solution The equation of the curve can be rewritten as 5␲
␪⫽
F(r, θ ) = r − a(1 + cos θ) = 0. The following observation can 3
be made about the curve: Fig. 8.18 Curve for r = a(1 + cos θ)
(a) The curve is symmetric about the initial line, as F(r, −θ ) =
F(r, θ ) 8.9.3 Parametric Equations
(b) r = 0 for cos θ = −1, i.e., θ = π . Hence, the pole lies on
the curve. Parametric equations of a plane curve may be written as
(c) The curve has no asymptote, as |r | = |a(1 + cos θ )| = x = f (t), y = g(t), t being a parameter. Sometimes para-
2|a| cos2 θ2 ≤ 2a for every value of θ. metric equations may be reduced to a Cartesian equation
(d) Putting θ = 0 in the equation of the curve and solving for by eliminating the parameter t between x = f (t) and
r , we find that the point of intersection of the curve with y = g(t). In that case, the characteristics of the curve as
the initial line is (2a, 0). discussed in Section 8.9.1 may be used for curve tracing.
(e) Since |r | = |a(1 + cos θ)| ≤ 2a, curve lies within the circle Let us, therefore, consider in this section the cases where
with radius 2a and with centre at the pole. parameter t cannot be eliminated.
192 Engineering Mathematics

Symmetry x(t) → 2a and y(t) → ∞. Hence, x = 2a is an asymptote

If x = f (t) is an even function, whereas y = g(t) an odd to the curve.
function, then the curve is symmetric about the x-axis.
Similarly, if y = g(t) is an even function, whereas x = f (t)
an odd function, then the curve is symmetric about the
y-axis. Y
2at2 , 2at3
x⫽ y⫽
Illustration 8.32 The curve defined by x = t 2 , y = 4t t2 ⫹ 1 t2 ⫹ 1
is symmetric about the x-axis, whereas the curve defined x⫽2a
by x = 2t, y = t 2 − 2 is symmetric about the y-axis.

Origin O (0, 0) X
Solve the equations f (t) = 0, g(t) = 0 for t. If for any value
of t, both the equations f (t) = 0 and g(t) = 0 are satisfied,
then the origin lies on the curve.

Illustration 8.33 The curve x = t 2 , y = 4t passes

through the origin, whereas the curve defined by x = 2t, y = 2at 2
t 2 − 2 does not. Fig. 8.19 Asymptote to the parametric curve x = ,
t2 + 1
2at 3
y=
Intercepts t2 + 1
The x-intercepts can be obtained by putting those values
of t in x = f (t), for which g(t) = 0. Similarly, y-intercepts
can be obtained by putting those values of t in y = g(t), Tangent
for which f (t) = 0. The gradient at any point t on the curve is given by

Illustration 8.34 The x-intercepts of the curve dy dy/dt g  (t)

√ √ = = 
dx dy/dt f (t)
x = 2t, y = t 2 − 2 are x = ±2 2, as y = 0 for t = ± 2.
g  (t)
Thus, the tangent at any point is horizontal if = 0
Extent f  (t)
f  (t)
Extent of the curve in the x-direction can be found from the and it is vertical if = 0.
g  (t)
least and the greatest values of function x = f (t). Similarly,
the extent in the y-direction can be found from the least Example 8.25 Sketch the curve x = a (t + sin t),
and the greatest values of y = g(t). y = a (1 + cos t), t being a parameter.
Curve exists only for those values of t for which x = f (t)
and y = g(t) are real. Solution The following observation can be made about the
given curve.
Illustration 8.35 The extent of the curve x = 2t, y =
(a) Since y = a(1 + cos t) is an even function and x = a(t +
t 2 − 2 in the x-direction is −∞ to +∞. The extent of the
sin t) is an odd function, the curve is symmetric about the
curve in the y-direction is −2 to +∞.
y-axis.
(b) The origin does not lie on the curve.
Asymptotes (c) Here y = a(1 + cos t) = 0 gives t = ±π, ±3π, ±5π, . . .,
If there exists a value t1 such that lim f (t) = ∞ and/or hence the x-intercepts of the curve are x = ±aπ, ±3aπ,
t →t1
lim g(t) = ∞, then t = t1 is an asymptote. ±5aπ, . . ..
t → t1
x = a (t + sin t) = 0 implies t = 0, so the y-intercept of
Illustration 8.36 Consider the parametric equations the curve is a(1 + cos 0) = 2a.
(d) The least and the greatest values of function x = a (t +
2at 2 2at 3 sin t) are −∞ and ∞, respectively. The curve is extended
x= ,y= 2
t +1
2 t +1 from −∞ to ∞ in the x-direction.
The least and the greatest values of function y = a(1 +
The curve described by these equations is known as cis-
cos x) are 0 and 2a, respectively. Hence, the curve is ex-
soid and is shown in Fig. 8.19. Observe that as t → ∞,
tended between the x-axis and the horizontal line y = 2a.
 Applications of Differential Calculus 193

(e) For any finite value of t, the x and y values on the curve (g) For different values of t, values of x and y are as fol-
are finite, hence the curve has no asymptote. lows.
(f) The gradient at any point t on the curve is given by
dy dy/dt −a sin t t t x y t x y
= = = − tan
dx d x/dt a(1 + cos t) 2
t 0 0.00 2.00 a π 3.14 a 0.00
Thus, the tangent at any point is horizontal if tan = 0, π 7π
2 1.02 a 1.87 a 3.17 a 0.13 a
i.e., if t = 0, ±2π, ±4π, . . .. 6 6
t
The tangent is vertical if cot = 0, i.e., if t = ±π, ±3π, π 5π
2 1.49 a 1.71 a 3.22 a 0.29 a
±5π, . . .. 4 4
π 4π
1.91 a 1.50 a 3.32 a 0.50 a
y 3 3
2.0a π 3π
2.57 a 1.00 a 3.71 a 1.00 a
2 2
1.5a 2π 5π
2.96 a 0.50 a 4.37 a 1.50 a
3 3
1.0a 3π 7π
3.06 a 0.29 a 4.79 a 1.71 a
4 4
0.5a 5π 11π
3.12 a 0.13 a 5.26 a 1.87 a
6 6
x π 3.14 a 0.00 2π 6.29 a 2.00 a
0 2a 3.14a 4a 6a
From all the above information, we can draw the curve as in
Fig. 8.20 Curve of x = a(t + sin t), y = a(1 + cos t)
Fig. 8.20.

Section Review 8.8

1. Sketch, approximately, the curve described by the equa- 5. Sketch the lemniscate r 2 = a 2 sin 2θ, where a is a positive
tion y = (x − 2)(x + 2)(x − 5). number.
x2 − 4 6. Draw the polar curves (a) r = 2 cos 2θ and
2. Draw the curve whose equation is y = 2 .
x +4 (b) r = 3 sin 3θ
3. Find an approximate sketch of curve described by
7. Draw the parametric curves.
x 3 − 5x + 1 (a) x = a(t + sin t), y = a(1 − cos t)
y= .
x2 − 4
√ 3at 3at 2
x +1 (b) x = ,y= (a > 0)
4. Sketch the curve whose equation is y = . 1+t 3 1 + t3
x +2

8.10 Maxima and Minima

Let f (x, y, z, . . .) be a function of several independent vari- To find (a, b, c, . . .), the values of (x, y, z, . . .) for which
ables x, y, z, . . . and let it be continuous and finite for all the function attains a maximum or a minimum, let us ex-
values of these variables in some neighbourhood of the point pand the function by Taylor’s theorem, i.e.
(a, b, c, . . .). Then, the function is said to have a maxi- f (x + h, y + k, z + l, . . .)
mum at (a, b, c, . . .) if f (x, y, z, . . .) < f (a, b, c, . . .) for ∂f ∂f ∂f
all (x, y, z, . . .) in the neighbourhood of (a, b, c, . . .), and = f (x, y, z, . . .) + h +k +l + · · · + R2
∂x ∂y ∂z
(x, y, z, . . .) = (a, b, c, . . .). or f (x + h, y + k, z + l, . . .) − f (x, y, z, . . .)
Similarly, the function is said to have a minimum at ∂f ∂f ∂f
(a, b, c, . . .) if f (x, y, z, . . .) > f (a, b, c, . . .) for all =h +k +l + · · · + R2
∂x ∂y ∂z
(x, y, z, . . .) in the neighbourhood of (a, b, c, . . .), and where R2 consists of second-and higher-power terms of
(x, y, z, . . .) =
 (a, b, c, . . .). h, k, l, . . ..
194 Engineering Mathematics

For sufficiently small h, k, l, . . ., the first-degree terms Solving these two equations, we obtain x = 30, y =
∂f ∂f ∂f −20. (In Illustration 8.40 it is observed that the func-
h +k +l + ··· tion has neither a minimum nor a maximum at the
∂x ∂y ∂z
point x = 30, y = −20.)
govern the sign of the right-hand side, and therefore, of the ∂f ∂f
left-hand side of the above equation. Thus, the conditions = 0 and = 0 are not suffi-
∂x ∂y
It is clear that the sign of the left-hand side will change cient for a maximum or a minimum.
with change of the signs of h, k, l, . . ..
Thus, for a maximum or a minimum of the function 8.10.1 Critical and Saddle Points
f (x, y, z, . . .), we must have
Critical Point A point (a, b) is said to be a critical
∂f ∂f ∂f ∂f
h +k +l + ··· = 0 point of the function f (x, y) if f (a, b) = 0, (a, b) = 0,
∂x ∂y ∂z ∂x
∂f
Since h, k, l, . . . are independent of each other, we must and (a, b) = 0.
∂y
have
∂f ∂f ∂f Figures 8.21 and 8.22 show critical points for a function
= 0, = 0, = 0, . . . (8.26) f (x, y). Fig. 8.21 shows the critical point when a function
∂x ∂y ∂z
If f is a function of n independent variables, then Eqn is minimum at the point P (a, b). Similarly, Fig. 8.22 shows
(8.26) is a system of n simultaneous equations. By solving the critical point when a function is maximum at the point
this system of n simultaneous equations, we shall find the P (a, b).
possible values of x, y, z, . . . for which the function f may
attain one of its maxima or minima.
Thus, Eqn (8.26) forms a set of necessary conditions for
a maximum or a minimum of the function. However, it
should be noted that as these equations are not sufficient, f(x, y)
further tests have to be employed to make any conclusion.
Let us now consider the cases of functions with two vari- P(a, b)
ables. Assume that a, b is a set of values of x, y satisfying
Eqn (8.26).
Illustration 8.37 Consider the functions (a) f (x, y) = Fig. 8.21 Critical point when the function has a mini-
2 − x 2 − y 2 , (b) f (x, y) = 16x + 12y + x 2 + y 2 , and mum at the point P(a, b)
(c) f (x, y) = 20x + 40y − x 2 + 4x y − 2y 2 .
(a) Taking the first partial derivatives, we have
∂f ∂f P(a, b)
= −2x, = −2y
∂x ∂y
∂f ∂f f(x, y)
Putting = 0 and = 0 gives x = 0, y = 0, respec-
∂x ∂y
tively. (Later in Illustration 8.39 it is observed that the
function has a maximum at the point (0, 0).)
(b) Taking the first partial derivatives, we obtain Fig. 8.22 Critical point when the function has a maxi-
∂f ∂f mum at the point P(a, b)
= 16 + 2x and = 12 + 2y
∂x ∂y
∂f ∂f
Solving = 0 and = 0, we get x = −8, y = −6. Saddle Point A critical point (a, b) of a function f (x, y)
∂x ∂y
(Later in Illustration 8.39 it is observed that the func- is said to be a saddle point if f (x, y) has neither a maximum
tion has a minimum at the point (−8, −6).) nor a minimum at (a, b).
(c) Taking the first partial derivatives, we obtain Illustration 8.38 Consider the function f (x, y) = x 2 −y 2
∂f ∂f and the point on the surface corresponding to x = 0, y = 0.
= 20 − 2x + 4y and = 40 + 4x − 4y
∂x ∂y We see that f x (0, 0) = 0 and f y (0, 0) = 0. Hence, (0, 0)
∂f ∂f is a critical point of the function. However, the function
Now = 0 and = 0 gives 20 − 2x + 4y = 0 and has neither a maximum nor a minimum at the point (0, 0);
∂x ∂y
40 + 4x − 4y = 0 respectively. (0, 0) is a saddle point of the function as shown in Fig. 8.23.
 Applications of Differential Calculus 195

A Thus, when r t = s 2 , we need further investigation to

B find if the function has an optimum at the point. We can
check the values of the function in the neighbourhood of
the point. Otherwise, we have to consider again the terms
f(0, 0) of higher degrees. The cubic terms must vanish collectively
and the biquadratic terms must collectively be always (in-
D dependent of sign of h and k) either positive (for a mini-
mum) or negative (for a maximum).
f(x, y)⫽x2 ⫺y2
Illustration 8.39 Consider the function f (x, y) =
2 − x 2 − y2.
∂f
Taking the first partial derivatives, we have = −2x and
C ∂x
∂f
= −2y.
Fig. 8.23 Saddle point of the function f (x, y) = x 2 − y 2 ∂y
Taking the second partial derivatives, we have
8.10.2 Lagrange’s Conditions ∂2 f ∂2 f ∂2 f
r= = −2, s = = 0, and t = = −2
If f (x, y) has a maximum or a minimum at (a, b), then ∂x2 ∂ x∂ y ∂ y2
∂f ∂f ∂f ∂f
= 0, =0 = 0 and = 0 give x = 0, y = 0, respectively.
∂x ∂y ∂x ∂y
Also, if Thus, (0, 0) is a critical point of the function.
Now, r t − s 2 = (−2) (−2) − (0) (0) = 4 > 0.
' ( ' ( ' (
∂2 f ∂2 f ∂2 f As r is negative and r t − s 2 is positive, hence, the func-
r= ,s= , and t =
∂x2 x=a,y=b ∂ x∂ y x=a,y=b ∂ y2 x=a,y=b tion has a maximum at the point (0, 0).
1 Now, consider another function f (x, y) = 16x + 12y +
then f (a + h, b + k) − f (a, b) = (r h 2 + 2shk + tk 2 ) + R3 , x 2 + y2.
2
where R3 stands for the terms containing the third and Taking the first partial derivatives, we obtain
higher degree of h and k.
∂f ∂f
Thus, when h and k are sufficiently small, the sign of = 16 + 2x and = 12 + 2y
∂x ∂y
the right side of the equation is governed by the sign of
r h 2 + 2shk + tk 2 . Therefore, if r h 2 + 2shk + tk 2 is negative Differentiating f (x, y) the second time partially w.r.t. x
for all such values of h and k, then we have a maximum and y, we have
at (a, b). On the other hand, if r h 2 + 2shk + tk 2 is positive
∂2 f ∂2 f ∂2 f
for all such values of h and k, then we have a minimum at r= = 2, s = = 0, and t = =2
∂x2 ∂ x∂ y ∂ y2
(a, b).
Now, we can write ∂f ∂f
Solving = 0 and = 0, we get x = −8 and y = −6,
1 k 2 ∂x ∂y
r h 2 + 2shk + tk 2 = (r h + sk)2 + (r t − s 2 ) respectively.
r r
Thus, we get the critical point at (−8, −6).
Therefore, r h 2 + 2shk + tk 2 is invariably negative if r is Now, at the point (−8, −6), r t −s 2 = (2) (−2)−(0)2 = 4 > 0.
negative and r t > s 2 ; and r h 2 + 2shk + tk 2 is invariably pos- Since r is positive and r t − s 2 is also positive, hence, the
itive if r is positive and r t > s 2 . function has a minimum at the point (−8, −6).
Thus, f (x, y) has a maximum at (a, b) if r is negative
and r t > s 2 ; it has a minimum at (a, b) if r are positive Illustration 8.40 Consider the function f (x, y) = 20x +
and r t > s 2 . 40y − x 2 + 4x y − 2y 2 .
If r t < s 2 , f (x, y) has neither a maximum or a minimum Taking the first partial derivatives, we have
at the point. ∂f ∂f
1 = 20 − 2x + 4y and = 40 + 4x − 4y
If r t = s 2 , r h 2 + 2shk + tk 2 becomes (hr + ks)2 , and ∂x ∂y
r
s h
therefore, has the same sign of r or t unless = − . Taking the second partial derivatives of f (x, y), we have
r k
However, if h and k are taken small but in the ratio
h s ∂2 f ∂2 f ∂2 f
= − , then r h 2 + 2shk + tk 2 becomes zero. r= = −2, s = = 4, and t = = −4
k r ∂x 2 ∂ x∂ y ∂ y2
196 Engineering Mathematics

∂f ∂f Taking the second partial derivatives, we have

Putting = 0 and = 0 gives 20 − 2x + 4y = 0 and
∂x ∂y
40 + 4x −4y = 0, respectively. ∂2 f ∂2 f ∂2 f
r= = 2, s = = −4y, and t = = −4x + 60y 2
Solving these two equations, we get x = −30, y = −20. ∂x2 ∂ x∂ y ∂ y2
Thus, the function has a critical point at (−30, −20). At the point (0, 0), we see that
Now, at the point (−30, −20), r t − s 2 = (−2)(−4) − (4)2
∂f ∂f
= −8 < 0. = 0 and =0
Hence, (−30, −20) is a saddle point of the function z = ∂x ∂y
f (x, y). Hence, the function may have an extremum at (0, 0).
Illustration 8.41 Consider the function f (x, y) = x − 2 At the point (0, 0), we have r = 2, s = 0, and t = 0, and
2x y 2 + 5y 4 and check the point (0, 0) for a minimum or hence, r t − s 2 = (2)(0) − (0)2 = 0.
maximum. Hence, nothing conclusive about the extremum can be
Taking the first partial derivatives, we have obtained from this result.
2
However, it can be observed that f (x, y) = x − y 2 +
∂f ∂f
= 2x − 2y 2 and = −4x y + 20y 3 . 4y ≥ 0 = f (0, 0) for all values of x and y. Thus, the
4
∂x ∂y
function f (x, y) has a minimum at the point (0, 0).

Section Review 8.9

x
1. Find the extrema of the following functions: 2. Show that the minimum value of is e.
log x
(a) f (x, y) = 4x y − x 2 − y 2 3. Find the point P(a, b, c) on the surface defined by
(b) f (x, y) = x 2 + 2y 2 − x y − 7x
(c) f (x, y) = x 2 − 2x y − y 2 + 10x − 6y 2
z = − y 2 − x 2 − x y + 3x
(d) f (x, y) = x 3 + y 3 − 3x − 12y + 12 3
(e) f (x, y) = x 3 + 3x y 2 − 3y 2 − 3x 2 + 4 with the greatest z co-ordinate.

8.11 Constrained Maxima and Minima

Let u = f (x1 , x2 , . . . , xn ) be a function of n variables and From the equation x + y = m, we have y = m − x and
suppose these variables are not all independent, rather they substituting this value of y in f (x, y), we have the function
are connected by the following m equations: log x + log (m − x) of single variable x for optimization.
f 1 (x1 , x2 , . . . , xn ) = 0 By Lagrange’s condition for an optimum, we have
f 2 (x1 , x2 , . . . , xn ) = 0 ' (
d 1 1
··· − =0
dx x m−x
f m (x1 , x2 , . . . , xn ) = 0
1 1 m
These m equations connecting the variables x1 , x2 , . . . , xn i.e., − = 0 or x =
x m−x 2
are generally called constraints. Assuming that these m
constraint equations are linearly independent, only (n − m) m m
Then, y = m − x = m − = .
of the n variables x1 , x2 , . . . , xn are independent. 2 2 m
To find maxima or minima, m of the n variables may The function f (x, y) has the optimal value when x =
m 2
be eliminated from u = f (x1 , x2 , . . . , xn ) with the help of and y = .
m constraint equations to make f a function of (n − m) 2
In Illustration 8.42, the elimination of one variable (y in
independent variables and then Lagrange’s conditions may
terms of x) is very easy. However, elimination of m variables
be used.
by expressing them in terms of the remaining (n − m) inde-
Illustration 8.42 Suppose we want to maximize f (x, y) = pendent variables is not always an easy process. Also, when
log x + log y subject to the constraint g(x, y) = x + y = m, m dependent variables are eliminated from the function to
where m being a constant. be optimized, to make it a function of (n − m) independent
Here, the two independent variables x and y are co- variables, the function sometimes takes a complicated form,
nnected by the constraint equation g(x, y) = x + y = m. and hence, Lagrange’s conditions give rise to complicated
One of these variables may be expressed in terms of the equations. For these reasons, elimination of dependent vari-
other, and hence, may be eliminated. ables are sometimes avoided.
 Applications of Differential Calculus 197

Lagrange’s multiplier method, another way for optimiza- Illustration 8.43 Consider again the optimization prob-
tion of function with constraint equations, does not require lem of Illustration 8.42 and solve it using Lagrange’s mul-
elimination of dependent variables of the function to be tiplier method.
optimized. Form the Lagrangian function as
L = log x + log y + λ(m − x − y)
8.11.1 Lagrange’s Multiplier Method Taking partial derivatives and equating to zero, we have
Let u = f (x1 , x2 , . . . , xn ) be a function, as defined earlier, ∂L 1
of n variables, m constraint, and only (n − m) independent = −λ=0 (8.27)
∂x x
variables.
∂L 1
The function u has a maximum or a minimum at a point = − −λ=0 (8.28)
∂y y
(x1 , x2 , . . . , xn ) if
and
∂f ∂f ∂f ∂L
du = d x1 + d x2 + . . . + d xn = 0 =m−x−y =0 (8.29)
∂ x1 ∂ x2 ∂ xn ∂λ
Eliminating λ from Eqn (8.27) and Eqn (8.28), we get x =
Also
y.
m
∂ f1 ∂ f1 ∂ f1 Hence, from Eqn (8.29), we get x = y = .
d f1 = d x1 + d x2 + . . . + d xn = 0, 2
∂ x1 ∂ x2 ∂ xn
2
Here, the value of multiplier λ is .
∂ f2 ∂ f2 ∂ f2 m
d f2 = d x1 + d x2 + . . . + d xn = 0, We get the same result by both the methods (Illustra-
∂ x1 ∂ x2 ∂ xn
tions 8.42 and 8.43).
...
∂ fm ∂ fm ∂ fm Example 8.26 Find the optimal value of the function
d fm = d x1 + d x2 + . . . + d xn = 0
∂ x1 ∂ x2 ∂ xn f (x, y) = x 2 −3x y +12x subject to the constraint 2x +3y = 6.
Multiplying these equations by 1, λ1 , λ2 , . . . , λm , respec- Solution Form the Lagrangian function as L = x 2 − 3x y +
tively, and then adding, we get 12x + λ(6 − 2x − 3y). Taking partial derivatives and equating
to zero, we have
P1 d x1 + P2 d x2 + P3 d x3 + . . . + Pn d xn = 0 ∂L
= 2x − 3y + 12 − 2λ = 0 (8.30)
∂x
∂f ∂ f1 ∂ f2 ∂ fm ∂L
where Pr = + λ1 + λ2 + . . . + λm = −3x − 3λ = 0 (8.31)
∂ xr ∂ xr ∂ xr ∂ xr ∂y
and
for r = 1, 2, . . . , n. ∂L
The quantities λ1 , λ2 , . . . , λm are arbitrary, hence, here = 6 − 2x − 3y = 0 (8.32)
∂λ
they are choosen such that Pr = 0 for r = 1, 2, . . . , m. Eliminating y from Eqn (8.30) and Eqn (8.32), we get
The above equation now reduces to
4x − 2λ + 6 = 0 (8.33)

Pm+1 d xm+1 + Pm+2 d xm+2 + · · · + Pn d xn = 0 Solving Eqn (8.31) and Eqn (8.33), we have λ = 1 and x = −1.
Finally substituting these values in Eqn (8.30), we have
8
It does not matter which (n − m) of the n variables are con- y= .
sidered as independent variables; so, assume that x m+1 , xm+2 , 3
Thus, the function f (x, y) = x 2 − 3x y + 12x subject to the
. . . , xn are independent. Since d xm+1 , d xm+2 , . . . , d xn are all
constraint 2x + 3y = 6 has an optimal value when x = −1 and
independent, the coefficients Pm+1 , Pm+2 , . . . , Pn must sep-
8
arately be zero. y= .
3
Thus, we obtain a set of additional (n − m) equations
Pm+1 = Pm+2 = . . . = Pn = 0. Example 8.27 A manufacturer can produce three distinct
From the (m + n) equations f 1 = f 2 = . . . = f m = 0 products in quantities q1 , q2 , and q3 , respectively, and thereby
and P1 = P2 = . . . = Pm = Pm+1 = . . . = Pn = 0, we can derive a profit p(q1 , q2 , q3 ) = 2q1 + 8q2 + 24q3 .
determine the m quantities λ1 , λ2 , . . . , λm and the values of Find the values of q1 , q2 , and q3 that maximize profit
the n variables for which maxima and minima values of the if production is subject to the constraint q12 + 2q22 + 4q32 =
function u = f (x1 , x2 , . . . , xn ) are possible. 4,500,000,000.
198 Engineering Mathematics

Solution The profit p(q1 , q2 , q3 ) = 2q1 + 8q2 + 24q3 has to i.e., 2 + λ 2q1 = 0, 8 + λ 4q2 = 0, and 24 + λ 8q3 = 0 together
be maximized subject to the constraint g(q1 , q2 , q3 ) = q12 + 2q22 with q12 + 2q22 + 4q32 − 4,500,000,000 = 0
+ 4q32 − 4,500,000,000 = 0. Solving q1 , q2 , and q3 in terms of λ, we have
For maximum value of profit p(q1 , q2 , q3 ), q1 , q2 , and q3 1 2 3
must satisfy Lagrange’s conditions, i.e., q1 = − , q2 = − , q3 = −
λ λ λ
∂ p(q1 , q2 , q3 ) ∂g(q1 , q2 , q3 ) Thus, we have
+λ = 0
∂q1 ∂q1 1 4 9
+ 2 2 + 4 2 − 4,500,000,000 = 0
λ2 λ λ
∂ p(q1 , q2 , q3 ) ∂g(q1 , q2 , q3 )
+λ = 0 1 1
∂q2 ∂q2 or λ2 = or λ = − [q1 , q2 , and q3
100,000,000 10,000
∂ p(q1 , q2 , q3 ) ∂g(q1 , q2 , q3 ) being all positive, λ must be negative.]
+λ = 0
∂q3 ∂q3 Hence, for maximum profit, we must have q1 = 10,000,
together with g(q1 , q2 , q3 ) = 0. q2 = 20,000, and q3 = 30,000.

Section Review 8.10

1. Find the point P(a, b, c) with the greatest z co-ordinate 9. Find the maximum and minimum values of
on the intersection of the surface defined by x2 y2 z2
+ +
2 a4 b4 c4
z = y 2 − x 2 − x y + 3x
3 x2 y2 z2
with the plane 2x − 3y = 0. when lx + my + nz = 0 and 2
+ 2 + 2 = 1.
a b c
2. Find the dimensions of the rectangular parallelopiped 10. Find the dimensions of the largest rectangular para-
of maximum volume which is located with of its faces llelopiped that has three faces in the co-ordinate planes
lying in the co-ordinate planes and with a vertex lying and vertex in plane
x y z
in the plane with equation + + = 1, where a, b, x y z
a b c + + =1
and c are positive real numbers. Also find this maximum a b c
volume. 11. Find the maxima and minima of x 2 + y 2 + z 2 subject to
3. Find the greatest value of u = x yz, if x, y, and z are the following conditions:
positive real numbers for which 4x + 2y + z = 12.
4. Find the minimum value of x 2 + y 2 + z 2 when it is given (a) ax 2 + by 2 + cz 2 = 1
that ax + by + cz = p. (b) ax 2 + by 2 + cz 2 + 2 f yz + 2gzx + 2 f x y = 1
5. Find the minimum value of x m + y n + z p subject to the (c) ax 2 + by 2 + cz 2 = 1 and lx + my + nx = 0
condition x + y + z = a. (d) ax 2 + by 2 + cz 2 + 2 f yz + 2gzx + 2 f x y = 1
6. Find the maximum value of cos A cos B cos C, where and lx +my + nx = 0
A + B + C = π. 12. Find the minimum value of the function f (x, y) =
7. A top-open rectangular box is required to be of capacity x 2 + (y − 2)2 on the hyperbola x 2 − y 2 = 1.
32 cubic ft. Find the dimensions of the box so that total 13. Show that of all the triangles inscribed in a fixed circle,
surface area is minimum. the equilateral triangle has the largest perimeter.
8. Divide a number n into three parts x, y, and z such that 14. Using Lagrange’s multiplier method, find the minimum
ayz + bzx + cx y shall have a maximum or a minimum and maximum distance of the point (4, 5) from the circle
value and determine that value. x 2 + y 2 = 4.

8.12 Application to Engineering Problems

There are many applications of integral calculus in real life use of transition curves to connect the straight part of
and engineering problems. A few of these applications are a track with a circular track. Generally, arcs of cubical
discussed here. parabolas are employed.
Suppose the transition curve on a railway track has
1. Transition Curve for a Rail Track At high speeds 1
trains cannot pass abruptly from a straight stretch of the shape of an arc of the cubical parabola y = x 3 ,
3
track to a circular track of high curvature. In order to where x and y denote the measurements in miles. Find
make the change of direction gradual, engineers make the rate of change of direction of a train when
 Applications of Differential Calculus 199

 
1 Thus, we have
it is passing through the point 1, on this
3 (ω − ω0 )2
track. X (ω) = X (ω0 )+ X  (ω0 )(ω−ω0 )+x  (ω0 ) +. . .
1 dy 2!
By differentiation of y = x 3 , we have = x 2 and
3 d x When ω is very near to ω0 , the terms with second and
d2 y higher powers of (ω − ω0 ) can be ignored in this expan-
= 2x.
dx2 sion.
Substituting these in Eqn (8.12), we have Thus
 
2x 1
K = 3 X (ω) = L + 2 (ω − ω0 ) = 2L(ω − ω0 ).
1 + x4 2 ω0 C
 
1 2 1 This gives an approximate expression of reactance near
At 1, , K = 3 = √ radian per mile. This is the the resonance frequency.
3 22 2
 ofchange of direction of the train at the given point
rate 3. Manufacturing problem A CD (compact disc) com-
1 pany manufactures blank CD of radius 5 cm. If the ra-
1, .
3 dius of a manufactured CD is off by more than 0.02
2. Electrical Engineering When a current of frequency cm, it is automatically rejected. The company needs to
ω is applied to a series R LC circuit, the net reactance know the variation in the area of the discs which pass
1
is given by X = ωL − . Find an approximate but inspection.
ωC
simple expression for X at frequencies near the resonant This actually refers to the error in the area of CD
frequency ω0 . which may pass inspection. Writing r for radius and A
1 for area of a CD, we have A = πr 2 .
At resonant frequency, we have ω0 L = .
ω0 C So, dA = 2πr dr .
1
Now, X = ωL − is function of ω. So, we can write Taking r = 5 cm and dr = 0.02 cm, we have d A =
ωC 44
it as X (ω) and expand it as a Taylor series about the × 5 × 0.02 = 0.6286 sq cm.
7
point ω = ω0 .

EXERCISES

Multiple Choice Questions

0.3
2
1. To integrate f (x)d x, where f (x) = x 2 e x , if only the (c) three maxima and two minima points
0 (d) one maximum point and one minimum point
first term of the Maclaurin series for f (x) is used as an
6. The number of asymptotes of the curve y = e−x + 1 on
2
approximate value of the function, then the approximate
value of the integral is (−∞, ∞) is
(a) 0.008 (b) 0.009 (c) 0.082 (d) 0.088 (a) infinity (b) two (c) one (d) zero
x3 7. The polar curve r = 5 cos 5θ has
2. If for |x| < 1, we write sin x ≈ x − , then the numerical
3!
error never exceeds (a) no loop (b) 5 loops (c) 10 loops (d) 2 loops
(a) 0.003 (b) 0.005 (c) 0.007 (d) 0.009 8. The point on the curve y 2 = 4x, which is nearest to the
3. The curve y = x 2 (3 − x) point (2, 1) is
(a) is symmetric about the y-axis (a) (1, 2) (b) (1, −2)
 (c)π(0,
3 0) (d) None of the these
(b) is symmetric about the x-axis 9. The coefficient of x − in the Taylor series of cos x
(c) has a loop (d) None of these π 4
about is
4. The figure formed by all the asymptotes of the curve 4
1 1 1 1
x 2 y 2 = c2 x 2 + y 2 is a (a) (b) − (c) √ (d) − √
12 12 6 2 6 2
(a) parallelogram (b) rectangle 10. The function f (x, y) has a maximum at a point, if at
(c) triangle (d) square that point
x2 2
− 5t + 4t (a) fx = fy = 0, fx x f yy − fx y
2
< 0, and fx x <0
5. The function f (x) = dt has
0 2 + et (b) fx = fy = 0, fx x f yy − fx y
2
< 0, and fx x >0
(a) two maxima and two minima points 2
(c) fx = fy = 0, fx x f yy − fx y > 0, and fx x <0
(b) two maxima and three minima points 2
(d) fx = fy = 0, fx x f yy − fx y > 0, and fx x >0
200 Engineering Mathematics

Review Questions
1. For a curve given by the Cartesian equation y = f (x), 6. Define circle of curvature of a curve at a point on it.
write the expression for the rate of change of arc- Explain how it may be found for a curve given by a
length w.r.t. the co-ordinate x. Also express the dif- Cartesian equation.
ferential arc-length of the curve. 7. What is an evolute? State the procedure for finding
2. How is the rate of change of arc-length of a polar curve evolute for a curve in Cartesian co-ordinates.
w.r.t. a polar co-ordinate expressed? Express also the 8. Define an envelope. How can the envelope for a given
differential arc-length. family of curves be found?
3. Define curvature of a curve at a point on it. How is it 9. What is a critical point? Define a saddle point. State
expressed in Cartesian and polar co-ordinate systems? the criteria for a critical point to be a maximum point,
4. What do you mean by radius of curvature? How is it a minimum point, and a saddle point.
related to curvature? 10. State Lagrange’s conditions for optimization.
5. Write the expressions for the radii of curvature for 11. Describe the Lagrange’s multiplier method for opti-
parametric curve and also for the curve expressed by mization with constrains.
implicit equations.

Answers to Multiple Choice Questions

1. (b) 2. (d) 3. (d) 4. (d) 5. (a) 6. (c) 7. (b) 8. (a) 9. (c) 10. (d)