MATHEMATICS

STRAIGHT LINE Case 2. Ellipse e < 1:

The general form of the equation is
Ax + By + C = 0
The standard form of the equation is
y = mx + b,
which is also known as the slope-intercept form.
The point-slope form is y y1 = m(x x1)
Given two points: slope, m = (y2 y1)/(x2 x1)
The angle between lines with slopes m1 and m2 is
= arctan [(m2 m1)/(1 + m2m1)]
Two lines are perpendicular if m1 = 1/m2 is the standard form of the equation. When h = k = 0,
e = 1 - _b 2/a 2i = c/a
The distance between two points is
Eccentricity:
d = _ y2 - y1i + ^ x2 - x1h
2 2
b = a 1 - e 2;
QUADRATIC EQUATION Focus: ( ae, 0); Directrix: x = a/e
ax2 + bx + c = 0 Case 3. Hyperbola e > 1:

b ! b 2 - 4ac
x = Roots = -
2a
CONIC SECTIONS

^ x - hh _ y - ki
2 2
- = 1; Center at (h, k)
a2 b2
is the standard form of the equation.When h = k = 0,
Eccentricity: e = 1 + _b 2/a 2i = c/a
e = eccentricity = cos /(cos ) b = a e 2 - 1;
[Note: X and Y , in the following cases, are translated axes.] Focus: ( ae, 0); Directrix: x = a/e
Case 1. Parabola e = 1:
Brink, R.W., A First Year of College Mathematics, D. Appleton-Century Co., Inc., 1937.

(y k)2 = 2p(x h); Center at (h, k)

is the standard form of the equation. When h = k = 0,
Focus: (p/2, 0); Directrix: x = p/2

MATHEMATICS 21
Case 4. Circle e = 0: If a2 + b2 c is positive, a circle, center (a, b).
(x h)2 + (y k)2 = r2; Center at (h, k) is the general If a2 + b2 c equals zero, a point at (a, b).
form of the equation with radius If a2 + b2 c is negative, locus is imaginary.

r = ^ x - hh + _ y - k i QUADRIC SURFACE (SPHERE)

2 2

The general form of the equation is

(x h)2 + (y k)2 + (z m)2 = r2

with center at (h, k, m).
In a three-dimensional space, the distance between two points
is
d = ^ x2 - x1h + _ y2 - y1i + ^ z2 - z1h
2 2 2

LOGARITHMS
The logarithm of x to the Base b is defined by
logb (x) = c, where bc = x
Special definitions for b = e or b = 10 are:
ln x, Base = e
Length of the tangent from a point. Using the general form log x, Base = 10
of the equation of a circle, the length of the tangent is found To change from one Base to another:
from logb x = (loga x)/(loga b)
t2 = (x h)2 + (y k)2 r2 e.g., ln x = (log10 x)/(log10 e) = 2.302585 (log10 x)
by substituting the coordinates of a point P(x,y) and the
Identities
coordinates of the center of the circle into the equation and
logb bn = n
computing.
log xc = c log x; xc = antilog (c log x)
log xy = log x + log y
logb b = 1; log 1 = 0
log x/y = log x log y

TRIGONOMETRY
Trigonometric functions are defined using a right triangle.
sin = y/r, cos = x/r
tan = y/x, cot = x/y
Conic Section Equation csc = r/y, sec = r/x
The general form of the conic section equation is
Ax2 + Bxy + Cy2 + Dx + Ey + F = 0
where not both A and C are zero.
If B2 AC < 0, an ellipse is defined.
Law of Sines
If B2 AC > 0, a hyperbola is defined. a = b = c
If B2 AC = 0, the conic is a parabola. sin A sin B sin C
If A = C and B = 0, a circle is defined. Law of Cosines
If A = B = C = 0, a straight line is defined. a2 = b2 + c2 2bc cos A
x2 + y2 + 2ax + 2by + c = 0 b2 = a2 + c2 2ac cos B
is the normal form of the conic section equation, if that conic c2 = a2 + b2 2ab cos C
section has a principal axis parallel to a coordinate axis.
h = a; k = b
Brink, R.W., A First Year of College Mathematics, D. Appleton-Century Co., Inc.,
r = a2 + b2 - c Englewood Cliffs, NJ, 1937.

22 MATHEMATICS
Identities Polar Coordinates
csc = 1/sin x = r cos ; y = r sin ; = arctan (y/x)
sec = 1/cos r = x + iy = x 2 + y 2
tan = sin /cos x + iy = r (cos + i sin ) = rei
cot = 1/tan [r1(cos 1 + i sin 1)][r2(cos 2 + i sin 2)] =
sin2 + cos2 = 1 r1r2[cos (1 + 2) + i sin (1 + 2)]
tan2 + 1 = sec2 (x + iy) = [r (cos + i sin )]n
n

cot2 + 1 = csc2 = rn(cos n + i sin n)

sin ( + ) = sin cos + cos sin r1 _cos i1 + i sin i1i
= r1 8cos _i1 - i2i + sin _i2iB
r
cos ( + ) = cos cos sin sin r2 _cos i2 + i sin i2i 2

sin 2 = 2 sin cos Eulers Identity

cos 2 = cos2 sin2 = 1 2 sin2 = 2 cos2 1 ei = cos + i sin
tan 2 = (2 tan )/(1 tan2) ei = cos i sin
cot 2 = (cot2 1)/(2 cot ) ii - ii ii - ii
cos i = e + e , sin i = e - e
tan ( + ) = (tan + tan )/(1 tan tan ) 2 2i
cot ( + ) = (cot cot 1)/(cot + cot ) Roots
sin ( ) = sin cos cos sin If k is any positive integer, any complex number
cos ( ) = cos cos + sin sin (other than zero) has k distinct roots. The k roots of r
(cos + i sin ) can be found by substituting successively
tan ( ) = (tan tan )/(1 + tan tan )
n = 0, 1, 2, ..., (k 1) in the formula
cot ( ) = (cot cot + 1)/(cot cot )
sin (/2) = ! ^1 - cos ah /2 w= k r <cos c i + n 360c m + i sin c i + n 360c mF
k k k k
cos (/2) = ! ^1 + cosh /2
Also, see Algebra of Complex Numbers in the
tan (/2) = ! ^1 - cos ah / ^1 + cos ah
ELECTRICAL AND COMPUTER ENGINEERING
cot (/2) = ! ^1 + cos ah / ^1 - cos ah section.
sin sin = (1/2)[cos ( ) cos ( + )]
MATRICES
cos cos = (1/2)[cos ( ) + cos ( + )] A matrix is an ordered rectangular array of numbers with
sin cos = (1/2)[sin ( + ) + sin ( )] m rows and n columns. The element aij refers to row i and
sin + sin = 2 sin (1/2)( + ) cos (1/2)( ) column j.
sin sin = 2 cos (1/2)( + ) sin (1/2)( )
Multiplication
cos + cos = 2 cos (1/2)( + ) cos (1/2)( ) If A = (aik) is an m n matrix and B = (bkj) is an n s matrix,
cos cos = 2 sin (1/2)( + ) sin (1/2)( ) the matrix product AB is an m s matrix

C = _cij i = d ! ailblj n
n
COMPLEX NUMBERS
l=1
Definition i = - 1
(a + ib) + (c + id) = (a + c) + i (b + d) where n is the common integer representing the number
of columns of A and the number of rows of B
(a + ib) (c + id) = (a c) + i (b d)
(l and k = 1, 2, , n).
(a + ib)(c + id) = (ac bd) + i (ad + bc)
a + ib = ^a + ibh ^c - id h = ^ac + bd h + i ^bc - ad h Addition
c + id ^c + id h ^c - id h c2 + d2 If A = (aij) and B = (bij) are two matrices of the same size
m n, the sum A + B is the m n matrix C = (cij) where
cij = aij + bij.

Identity
The matrix I = (aij) is a square n n identity matrix where
aii = 1 for i = 1, 2, ..., n and aij = 0 for i j.

Transpose
The matrix B is the transpose of the matrix A if each entry
bji in B is the same as the entry aij in A and conversely. In
equation form, the transpose is B = AT.

MATHEMATICS 23
Inverse Addition and subtraction:
The inverse B of a square n n matrix A is A + B = (ax + bx)i + (ay + by)j + (az + bz)k
adj ^ Ah A B = (ax bx)i + (ay by)j + (az bz)k
B = A- 1 = , where
A
adj(A) = adjoint of A (obtained by replacing AT elements with The dot product is a scalar product and represents the
their cofactors, see DETERMINANTS) and projection of B onto A times A . It is given by
A = determinant of A.
AB = axbx + ayby + azbz
Also, AA1 = A1A = I where I is the identity matrix.
= A B cos i = B : A
DETERMINANTS The cross product is a vector product of magnitude
A determinant of order n consists of n2 numbers, called the B A sin which is perpendicular to the plane containing
elements of the determinant, arranged in n rows and n columns A and B. The product is
and enclosed by two vertical lines.
i j k
In any determinant, the minor of a given element is the A # B = ax ay az =- B # A
determinant that remains after all of the elements are struck
bx by bz
out that lie in the same row and in the same column as the
given element. Consider an element which lies in the jth
column and the ith row. The cofactor of this element is the The sense of A B is determined by the right-hand rule.
value of the minor of the element (if i + j is even), and it is
A B = A B n sin , where
the negative of the value of the minor of the element
(if i + j is odd). n = unit vector perpendicular to the plane of A and B.
If n is greater than 1, the value of a determinant of order n is Gradient, Divergence, and Curl
the sum of the n products formed by multiplying each element dz = c 2 i + 2 j + 2 k m z
of some specified row (or column) by its cofactor. This sum 2x 2y 2z
is called the expansion of the determinant [according to the
elements of the specified row (or column)]. For a second-order d : V = c 2 i + 2 j + 2 k m : _V1i + V2j + V3k i
2x 2y 2z
determinant:
a1a2 d # V = c 2 i + 2 j + 2 k m # _V1i + V2j + V3k i
= a1b2 - a2b1 2x 2y 2z
b1b2
The Laplacian of a scalar function is
For a third-order determinant: 22 z 22 z 22 z
a1 a2 a3 d2 z = 2 + 2 +
2x 2y 2z 2
b1 b2 b3 = a1b2c3 + a2b3c1 + a3b1c2 - a3b2c1 - a2b1c3 - a1b3c2
c1 c2 c3 Identities
A B = B A; A (B + C) = A B + A C
VECTORS A A = |A|2
ii=jj=kk=1
ij=jk=ki=0
If A B = 0, then either A = 0, B = 0, or A is perpendicular
to B.
A B = B A
A (B + C) = (A B) + (A C)
(B + C) A = (B A) + (C A)
ii=jj=kk=0
i j = k = j i; j k = i = k j
k i = j = i k
If A B = 0, then either A = 0, B = 0, or A is parallel to B.
d 2 z = d : ^dzh = ]d : dg z
d # dz = 0
d : ^d # Ah = 0
d # ^d # Ah = d ]d : Ag - d 2A
A = axi + ayj + azk

24 MATHEMATICS
PROGRESSIONS AND SERIES 4. Two power series may be added, subtracted, or multiplied,
Arithmetic Progression and the resulting series in each case is convergent, at least,
To determine whether a given finite sequence of numbers is in the interval common to the two series.
an arithmetic progression, subtract each number from the 5. Using the process of long division (as for polynomials),
following number. If the differences are equal, the series is two power series may be divided one by the other within
arithmetic. their common interval of convergence.
1. The first term is a. Taylors Series
2. The common difference is d.
f l ^ ah ^ f m ^ ah ^
3. The number of terms is n. f ^ xh = f ^ ah + x - ah + x - ah
2

4. The last or nth term is l. 1! 2!

^ nh
^ ah
5. The sum of n terms is S. ^ x - ah + ...
f n
+ ... + n!
l = a + (n 1)d
S = n(a + l)/2 = n [2a + (n 1) d]/2 is called Taylors series, and the function f (x) is said to be
Geometric Progression expanded about the point a in a Taylors series.
To determine whether a given finite sequence is a geometric If a = 0, the Taylors series equation becomes a Maclaurins
progression (G.P.), divide each number after the first by the series.
preceding number. If the quotients are equal, the series is
geometric. DIFFERENTIAL CALCULUS
1. The first term is a. The Derivative
2. The common ratio is r. For any function y = f (x),
3. The number of terms is n. the derivative = Dx y = dy/dx = y
4. The last or nth term is l.
5. The sum of n terms is S. y l = limit 8_Dy i / ^DxhB
Dx " 0
l = arn1
S = a (1 rn)/(1 r); r 1 = limit $7 f ^ x + Dxh = f ^ xhA / ^Dxh.
Dx " 0
S = (a rl)/(1 r); r 1 y l = the slope of the curve f (x) .
limit Sn= a/(1r); r < 1
n Test for a Maximum
y = f (x) is a maximum for
A G.P. converges if |r| < 1 and it diverges if |r| > 1. x = a, if f (a) = 0 and f (a) < 0.
Properties of Series Test for a Minimum
n
! c = nc; c = constant y = f (x) is a minimum for
i=1 x = a, if f (a) = 0 and f (a) > 0.
n n
! cxi = c ! xi Test for a Point of Inflection
i=1 i=1
y = f (x) has a point of inflection at x = a,
! _ xi + yi - zi i = ! xi + ! yi - ! zi
n n n n
if f (a) = 0, and
i=1 i=1 i=1 i=1 if f (x) changes sign as x increases through
! x = _n + n 2i /2 x = a.
n

x=1
Power Series
! i3= 0 ai ^ x - ah
i
1. A power series, which is convergent in the interval
R < x < R, defines a function of x that is continuous for
all values of x within the interval and is said to represent
the function in that interval.
2. A power series may be differentiated term by term within
its interval of convergence. The resulting series has the
same interval of convergence as the original series
(except possibly at the end points of the series).
3. A power series may be integrated term by term provided
the limits of integration are within the interval of
convergence of the series.
The Partial Derivative
In a function of two independent variables x and y, a
derivative with respect to one of the variables may be found
if the other variable is assumed to remain constant. If y is kept
fixed, the function
z = f (x, y)
becomes a function of the single variable x, and its derivative
(if it exists) can be found. This derivative is called the partial
derivative of z with respect to x. The partial derivative with
respect to x is denoted as follows:
2z 2f _ x, y i
2x = 2x

MATHEMATICS 25
The Curvature of Any Curve LHospitals Rule (LHpitals Rule)
If the fractional function f(x)/g(x) assumes one of the
indeterminate forms 0/0 or / (where is finite or infinite),
then
limit f ^ xh /g ^ xh
x " a

is equal to the first of the expressions

f l ^ xh f m ^ xh f n ^ xh
limit , limit , limit
x " a g l ^ xh x " a g m ^ xh x " a g n ^ xh

which is not indeterminate, provided such first indicated

limit exists.

The curvature K of a curve at P is the limit of its average INTEGRAL CALCULUS

curvature for the arc PQ as Q approaches P. This is also The definite integral is defined as:
limit ! f _ xi i Dxi = #a f ^ xh dx
n b
expressed as: the curvature of a curve at a given point is the
n " 3i=1
rate-of-change of its inclination with respect to its arc length.
Also, xi 0 for all i.
K = limit Da = da
Ds " 0 Ds ds
A table of derivatives and integrals is available in the
Curvature in Rectangular Coordinates Derivatives and Indefinite Integrals section. The integral
ym equations can be used along with the following methods of
K=
91 + _ y li2C
3 2 integration:
A. Integration by Parts (integral equation #6),
When it may be easier to differentiate the function with B. Integration by Substitution, and
respect to y rather than x, the notation x will be used for the C. Separation of Rational Fractions into Partial Fractions.
derivative.
x l = dx/dy
Wade, Thomas L., Calculus, Ginn & Company/Simon & Schuster Publishers, 1953.
K= - xm
81 + ^ x lh2B
3 2

The Radius of Curvature

The radius of curvature R at any point on a curve is defined as
the absolute value of the reciprocal of the curvature K at that
point.
R= 1 ^ K ! 0h
K
91 + _ y li2C
3 2

R= _ y m ! 0i
ym

26 MATHEMATICS
DERIVATIVES AND INDEFINITE INTEGRALS
In these formulas, u, v, and w represent functions of x. Also, a, c, and n represent constants. All arguments of the trigonometric
functions are in radians. A constant of integration should be added to the integrals. To avoid terminology difficulty, the
following definitions are followed: arcsin u = sin1 u, (sin u)1 = 1/sin u.
1. # d f (x) = f (x)
1. dc/dx = 0
2. # dx = x
2. dx/dx = 1
3. # a f(x) dx = a # f(x) dx
3. d(cu)/dx = c du/dx
4. # [u(x) v(x)] dx = # u(x) dx # v(x) dx
4. d(u + v w)/dx = du/dx + dv/dx dw/dx
^m ! - 1h
m+1
5. d(uv)/dx = u dv/dx + v du/dx 5. # x mdx = x
m+1
6. d(uvw)/dx = uv dw/dx + uw dv/dx + vw du/dx
d ^u/vh
6. # u(x) dv(x) = u(x) v(x) # v (x) du(x)
7. = v du/dx -2 u dv/dx
dx v 7. # dx = a1 1n ax + b
n n1
ax + b
8. d(u )/dx = nu du/dx
8. # dx = 2 x
9. d[f (u)]/dx = {d[f (u)]/du} du/dx x
10. du/dx = 1/(dx/du) x
d _logaui 9. # a x dx = a
= _logaei u1 du
11. 1na
dx dx 10. # sin x dx = cos x
^ h
12. d 1n u = u1 du 11. # cos x dx = sin x
dx dx
_ ui 12. # sin 2xdx = x - sin 2x
13. d a = ^1n ah au du 2 4
dx dx
u u
2
13. # cos xdx = + x sin 2x
14. d(e )/dx = e du/dx 2 4
15. d(uv)/dx = vuv1 du/dx + (ln u) uv dv/dx 14. # x sin x dx = sin x x cos x
16. d(sin u)/dx = cos u du/dx 15. # x cos x dx = cos x + x sin x
17. d(cos u)/dx = sin u du/dx 16. # sin x cos x dx = (sin2x)/2
18. d(tan u)/dx = sec2u du/dx 17. # sin ax cos bx dx =- cos ^a - bh x - cos ^a + bh x _a 2 ! b 2i
19. d(cot u)/dx = csc2u du/dx 2 ^a - bh 2 ^a + bh
20. d(sec u)/dx = sec u tan u du/dx 18. # tan x dx = lncos x= ln sec x
21. d(csc u)/dx = csc u cot u du/dx 19. # cot x dx = ln csc x = ln sin x
_ -1 i 20. # tan2x dx = tan x x
22. d sin u = 1 du _- r/2 # sin- 1u # r/2i
dx 1-u 2 dx 21. # cot2x dx = cot x x
22. # eax dx = (1/a) eax
_ -1 i
23. d cos u =- 1 du _0 # cos- 1u # ri 23. # xeax dx = (eax/a2)(ax 1)
dx 1-u 2 dx

_ tan- 1ui 24. # ln x dx = x [ln (x) 1] (x > 0)

24. d
= 1 du _- r/2 < tan- 1u < r/2i
dx 1 + u 2 dx 25. # dx = 1 tan- 1 x ^a ! 0h
_ -1 i a2 + x2 a a
25. d cot u =- 1 2 du _0 < cot u < ri
-1

= 1 tan- 1 b x ac l, ^a > 0, c > 0h

dx 1 + u dx dx
26. #
_ -1 i 2
ax + c ac
26. d sec u = 1 du
dx u u 2 - 1 dx
27a. # dx = 2 tan- 1 2ax + b 2
_0 < sec- 1u < r/2i _- r # sec- 1u < - r/2i ax 2 + bx + c 4ac - b 2 4ac - b
d _csc- 1ui 1 du _ 4ac - b 2 > 0i
27. =-
dx u u - 1 dx
2
2ax + b - b 2 - 4ac
_0 < csc u # r/2i _- r < csc u # - r/2i
dx 1
-1 -1 27b. # 2 = 1 n
ax + bx + c b 2 - 4ac 2ax + b + b 2 - 4ac
_b 2 - 4ac > 0i

27c. # dx =- 2 , _b 2 - 4ac = 0i
ax 2 + bx + c 2ax + b

MATHEMATICS 27
MENSURATION OF AREAS AND VOLUMES Circular Sector
Nomenclature
A = total surface area
P = perimeter
V = volume
Parabola

A = zr 2/2 = sr/2
z = s/r

Sphere

Ellipse

V = 4rr3/3 = rd3/6
A = 4rr 2 = rd 2

Parallelogram

Papprox = 2r _a 2 + b 2i /2
R V
S1 + _ 1 2i m 2 + _ 1 2 # 1 4i m 4
2 2
W
S W
P = r ^a + bh S+ _ 1 2 # 1 4 # 3 6 i m6 + _ 1 2 # 1 4 # 3 6 # 5 i m8 W,
2 2
8
S W
SS+ _ 1 # 1 # 3 # 5 # 7 i2 m10 + f
2 4 6 8 10
WW P = 2 ^a + bh
d1 = a 2 + b 2 - 2ab ^cos zh
T X
where
m = ^a - bh / ^a + bh d2 = a 2 + b 2 + 2ab ^cos zh
d12 + d22 = 2 _a 2 + b 2i
Circular Segment
A = ah = ab ^sin zh

A If a = b, the parallelogram is a rhombus.

Gieck, K. & Gieck R., Engineering Formulas, 6th ed., Gieck Publishing, 1967.

A = 8 r 2 ^z - sin zhB /2
z = s/r = 2 $arccos 7^r - d h /r A.

28 MATHEMATICS
MENSURATION OF AREAS AND VOLUMES Right Circular Cone
(continued)
Regular Polygon (n equal sides)

V = _rr 2hi /3
z = 2r/n A = side area + base area
^ h = rr ` r + r 2 + h 2 j
i = ; r n n- 2 E = r b1 - n2 l
P = ns Ax: Ab = x 2: h 2

s = 2r 8 tan ^z/2hB
Right Circular Cylinder
A = ^nsr h /2

Prismoid

2
V = rr 2 h = r d h
4
A = side area + end areas = 2rr ^h + r h

V = ^h/6h ^ A1 + A2 + 4Ah
Paraboloid of Revolution

2
V = rd h
8

Gieck, K. & R. Gieck, Engineering Formulas, 6th ed., Gieck Publishing, 1967.

MATHEMATICS 29
CENTROIDS AND MOMENTS OF INERTIA f(x) yp(x)
The location of the centroid of an area, bounded by the axes A B
and the function y = f(x), can be found by integration. Aex Bex, rn
# xdA A1 sin x + A2 cos x B1 sin x + B2 cos x
xc =
A If the independent variable is time t, then transient dynamic
# xdA solutions are implied.
yc =
A
A = # f ^ xh dx First-Order Linear Homogeneous Differential Equations
with Constant Coefficients
dA = f ^ xh dx = g _ y i dy y+ ay = 0, where a is a real constant:
Solution, y = Ceat
The first moment of area with respect to the y-axis and the where C = a constant that satisfies the initial conditions.
x-axis, respectively, are:
My = x dA = xc A First-Order Linear Nonhomogeneous Differential
Equations
Mx = y dA = yc A
+ y = Kx ^t h x ^t h = ) 3
dy A t<0
The moment of inertia (second moment of area) with respect x
dt B t>0
to the y-axis and the x-axis, respectively, are:
y ^0h = KA
Iy = x2 dA
Ix = y2 dA is the time constant
The moment of inertia taken with respect to an axis passing K is the gain
through the areas centroid is the centroidal moment of inertia. The solution is
The parallel axis theorem for the moment of inertia with y ^t h = KA + ] KB - KAg c1 - exp b -xt lm or
respect to another axis parallel with and located d units from
the centroidal axis is expressed by
x = 1n < KB - y F
t KB - KA
Iparallel axis = Ic + Ad2
In a plane, J =r2dA = Ix + Iy
Second-Order Linear Homogeneous Differential Equations
Values for standard shapes are presented in tables in the with Constant Coefficients
STATICS and DYNAMICS sections. An equation of the form
y+ ay+ by = 0
DIFFERENTIAL EQUATIONS
can be solved by the method of undetermined coefficients
A common class of ordinary linear differential equations is where a solution of the form y = Cerx is sought. Substitution of
d n y ^ xh dy ^ xh this solution gives
bn n + f + b1 + b0 y ^ xh = f ^ xh (r2 + ar + b) Cerx = 0
dx dx
and since Cerx cannot be zero, the characteristic equation must
where bn, , bi, , b1, b0 are constants. vanish or
When the equation is a homogeneous differential equation, r2 + ar + b = 0
f(x) = 0, the solution is The roots of the characteristic equation are
yh ^ xh = C1e r i x + C2e r 2 x + f + Cie r i x + f + Cne r n x r1, 2 =-
a ! a 2 - 4b
2
where rn is the nth distinct root of the characteristic and can be real and distinct for a2 > 4b, real and equal for
polynomial P(x) with a2 = 4b, and complex for a2 < 4b.
P(r) = bnrn + bn1rn1 + + b1r + b0 If a2 > 4b, the solution is of the form (overdamped)
If the root r1 = r2, then C2e r2 x
is replaced with C2xer1x. y = C1er1x + C2er2 x
Higher orders of multiplicity imply higher powers of x. The If a2 = 4b, the solution is of the form (critically damped)
complete solution for the differential equation is y = (C1+ C2x)er1x
y(x) = yh(x) + yp(x), If a2 < 4b, the solution is of the form (underdamped)
where yp(x) is any solution with f(x) present. If f(x) has ern x y = ex (C1 cos x + C2 sin x), where
terms, then resonance is manifested. Furthermore, specific f(x) = a/2
forms result in specific yp(x) forms, some of which are: 4b - a 2
b=
2

30 MATHEMATICS
FOURIER TRANSFORM NUMERICAL METHODS
The Fourier transform pair, one form of which is Newtons Method for Root Extraction
F ^~h = #- 3 f ^t h e- j~t dt
3
Given a function f(x) which has a simple root of f(x) = 0 at
f ^t h = 71/ ^2rhA #- 3 F ^~h e j~t d~
3 x = a an important computational task would be to find that
root. If f(x) has a continuous first derivative then the
(j +1)st estimate of the root is
can be used to characterize a broad class of signal models
in terms of their frequency or spectral content. Some useful
transform pairs are: f ^ xh
aj+1 = aj -
f(t) F() df ^ xh
dx x a j
d ^t h 1 =
u(t) rd ^~h + 1/j~ The initial estimate of the root a0 must be near enough to the
actual root to cause the algorithm to converge to the root.
^ h
u bt + x l - u bt - x l = rrect xt x sin ~x/2 Newtons Method of Minimization
2 2 ~x/2
Given a scalar value function
e j~o t
2rd _~ - ~oi
h(x) = h(x1, x2, , xn)
find a vector x*Rn such that
Some mathematical liberties are required to obtain the second h(x*) h(x) for all x
and fourth form. Other Fourier transforms are derivable from Newtons algorithm is
the Laplace transform by replacing s with j provided J N- 1
f ^t h = 0, t < 0
K 2 O
xk + 1 = xk - K 2 h2 O 2h , where
O 2x
f ^t h dt < 3
3 K 2x
#0 L x = xk P x = xk
R V
2
S W h
Also refer to Fourier Series and Laplace Transforms in the S 2x1 W
ELECTRICAL AND COMPUTER ENGINEERING S 2h W
S 2x W
section of this handbook. 2h = S 2 W
2x Sg W
DIFFERENCE EQUATIONS S W
Difference equations are used to model discrete systems. Sg W
S 2h W
Systems which can be described by difference equations S 2x W
include computer program variables iteratively evaluated in T X
n

a loop, sequential circuits, cash flows, recursive processes, and

systems with time-delay components, etc. Any system whose R 2 V
S2 h 2 2h g g 2 2h W
input v(t) and output y(t) are defined only at the equally spaced S 2x12 2x1 2x2 2x1 2xn W
intervals t = kT can be described by a difference equation. S W
S 2 2h 2 2h 2 2
h W
g g
S 2x 2x 2x22 2x2 2xn W
First-Order Linear Difference Equation 2 2h = S 1 2 W
The difference equation 2x 2 Sg g g g g W
S W
Pk = Pk1(1 + i) A Sg g g g g W
S 2 W
represents the balance P of a loan after the kth payment A. S 2h 2h 2
2 2h W
g g
If Pk is defined as y(k), the model becomes S 2x1 2xn 2x2 2xn 2xn2 W
y(k) (1 + i) y(k 1) = A T X
Second-Order Linear Difference Equation
The Fibonacci number sequence can be generated by
y(k) = y(k 1) + y(k 2)
where y(1) = 1 and y(2) = 1. An alternate form for this
model is f (k + 2) = f (k + 1) + f (k) with f (0) = 1 and f (1) = 1.

MATHEMATICS 31
Numerical Integration
Three of the more common numerical integration algorithms used to evaluate the integral
#a f ^ xh dx
b

are:
Eulers or Forward Rectangular Rule
#a f ^ xh dx . Dx ! f ^a + kDxh
b n-1

k=0

Trapezoidal Rule
for n = 1
f ^ah + f ^bh F
#a f ^ xh dx . Dx <
b
2
for n > 1
#a f ^ xh dx . Dx < f ^ah + 2 ! f ^a + kDxh + f ^bhF
b n-1

2 k=1

Simpsons Rule/Parabolic Rule (n must be an even integer)

for n = 2
#a f ^ xh dx . b b - a l ; f ^ah + 4f b a + b l + f ^bhE
b
6 2
for n 4
R V
S f ^ ah + 2 ! f ^a + kDxh W
n-2

S k = 2, 4, 6, f W
b
^ h
#a f x dx . D x S W
3 S W
^ h ^ h
n-1
S 4 f a + kDx + f b WW
S+ k = 1! , 3, 5, f
T X
with x = (b a)/n
n = number of intervals between data points

Numerical Solution of Ordinary Differential Equations

Eulers Approximation
Given a differential equation
dx/dt = f (x, t) with x(0) = xo
At some general time kt
x[(k + 1)t] x(kt) + tf [x(kt), kt]
which can be used with starting condition xo to solve recursively for x(t), x(2t), , x(nt).
The method can be extended to nth order differential equations by recasting them as n first-order equations.
In particular, when dx/dt = f (x)
x[(k + 1)t] x(kt) + tf [x(kt)]
which can be expressed as the recursive equation
xk + 1 = xk + t (dxk/dt)

Refer to the ELECTRICAL AND COMPUTER ENGINEERING section for additional information on Laplace transforms
and algebra of complex numbers.

32 MATHEMATICS