This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
0 0 < pq pq > 0
m = 1, 2,… m = 1, 2,…
DEFINITE INTEGRALS
∞
18.38.
∫
0
18.39.
∫
0
18.40.
∫
0
18.41.
∫
0
18.42.
∫
0
18.43.
∫
18.44.
∫
18.45.
∫
18.46.
∫
18.47.
∫
18.48.
∫
∞
∞
∞
∞
x sin mx π − ma e 2 2 dx = 2 x +a sin mx π dx = 2 (1 − e − ma ) x(x 2 + a2 ) 2a
0
π /2 0 2π 0 2π
0
∫
0
∫
0
18.52.
∫
0
18.53.
∫
0
18.54.
∫
18.55.
∫
2π a2 − b2
dx = a + b cos x
2π a2 − b2
dx cos −1 (b / a) = a + b cos x a2 − b2 dx = (a + b sin x )2
2π 0
dx 2π a = (a + b cos x )2 (a 2 − b 2 )3/2
dx 2π = , 1 − 2a cos x + a 2 1 − a 2
∞
∫
sin ax 2 dx =
0 < a <1
sin ax n dx =
∞
∞
∞ 0 ∞ 0 ∞
∫
cos ax 2 dx =
1 2
∞ 0
a <1 a >1 m = 0, 1, 2,…
π 2a n >1
1 π Γ (1/n) cos , 2n na1/n
cos ax n dx = sin x dx = x
∞ 0
a 2 < 1,
1 π Γ (1/n) sin , 2n na1/n
∞
0
∫
cos mx dx π am , 2 = 1 − 2a cos x + a 1 − a2
π
18.50.
dx = a + b sin x
⎧⎪(π /a) ln (1 + a), x sin x dx = 1 − 2a cos x + a 2 ⎨π ln (1 + 1/a), ⎪⎩
π
0
∫
cos mx π − ma e 2 2 dx = a 2 x +a
2π
∫
18.56.
cos px − cos qx π (q − p) dx = 2 2 x
0
18.49.
18.51.
cos px − cos qx q dx = ln x p
2π
0
111
n >1
cos x π dx = 2 x
sin x π dx = , 2Γ ( p) sin ( pπ /2) xp cos x π dx = , 2Γ ( p) cos ( pπ /2) xp
sin ax 2 cos 2bx dx =
1 2
π 2a
0 < p<1 0 < p<1
b2 b2 ⎞ ⎛ cos − sin ⎜⎝ a a ⎟⎠
DEFINITE INTEGRALS
112
18.57.
18.58.
∫ ∫
18.59.
∫
18.60.
∫
18.61.
∫
18.62.
∫
∞ 0 ∞ 0 ∞ 0 ∞ 0
cos ax 2 cos 2bx dx =
π 2a
b2 b2 ⎞ ⎛ cos sin + ⎜⎝ a a ⎟⎠
sin 3 x 3π dx = 8 x3 sin 4 x π dx = 3 x4 tan x π dx = x 2
π /2 0
π /2 0
dx π = 1 + tan m x 4
{
}
x 1 1 1 1 dx = 2 2 − 2 + 2 − 2 + sin x 1 3 5 7
tan −1 x 1 1 1 1 dx = 2 − 2 + 2 − 2 + x 1 3 5 7
1
∫
0
18.64.
∫
sin −1 x π dx = ln 2 0 2 x
18.65.
∫
0
18.66.
∫
18.67.
∫
18.63.
1 2
1
1
∞ cos x 1 − cos x dx − ∫ dx = γ 1 x x
∞ 0
∞ 0
⎛ 1 ⎞ dx ⎜⎝ 1 + x 2 − cos x⎟⎠ x = γ tan −1 px − tan −1 qx π p dx = ln x 2 q
Definite Integrals Involving Exponential Functions Some integrals contain Euler’s constant g = 0.5772156 . . . (see 1.3, page 3). ∞ a 18.68. ∫ e − ax cos bx dx = 2 0 a + b2 18.69.
∫
18.70.
∫
18.71.
18.72.
18.73.
∫
∞ 0
e − ax sin bx dx =
∞ − ax
e
0 ∞ − ax
e
0
∞
∫
0
∫
0
b sin bx dx = tan −1 x a − e − bx b dx = ln x a
e − ax dx =
∞
2
b a2 + b2
1 π 2 a
e − ax cos bx dx = 2
1 π − b /4 a e 2 a 2
DEFINITE INTEGRALS
18.74.
∫
∞ 0
2
∞
18.75.
∫
−∞
18.76.
∫
0
18.77.
∫
0
18.78.
∫
18.79.
∫
18.80.
∫
∞
∞
2
π
∫
∞
e − x dx
p
2
2
x n e − ax dx =
x m e − ax dx = 2
0 ∞ 0
2
2
π ( b − 4 ac )/4 a e a 2
Γ(n + 1) a n+1 Γ[(m + 1)/2] 2a ( m +1)/2
e − ( ax + b/x ) dx =
∞
2
e − ( ax + bx +c ) dx =
∞ 0
1 π ( b − 4 ac )/4 a b e erfc 2 a 2 a
e − ( ax + bx +c ) dx =
where erfc (p) =
113
1 π −2 e 2 a
ab
x dx π2 1 1 1 1 = + + + + = 6 e x − 1 12 22 32 4 2 x n−1 1 1 ⎛1 ⎞ dx = Γ(n) ⎜ n + n + n + ⎟ ex − 1 2 3 ⎝1 ⎠
For even n this can be summed in terms of Bernoulli numbers (see pages 142–143). 18.81. 18.82.
∫ ∫
∞ 0 ∞ 0
x dx π2 1 1 1 1 = 2 − 2 + 2 − 2 + = x 12 e +1 1 2 3 4 x n−1 1 1 ⎛1 ⎞ dx = Γ(n) ⎜ n − n + n − ⎟ x e +1 2 3 ⎝1 ⎠
For some positive integer values of n the series can be summed (see 23.10). 18.83.
∫
18.84.
∫
∞ 0
sin mx m 1 1 dx = coth − 4 2 2m e 2π x − 1
⎛ 1 − x ⎞ dx ⎜⎝ 1 + x − e ⎟⎠ x = γ
∞ 0
2 ∞ −x
e
− e− x dx = 12 γ x
18.85.
∫
18.86.
e− x ⎞ ⎛ 1 − ∫0 ⎜⎝ e x − 1 x ⎟⎠ dx = γ
18.87.
∫
18.88.
∫
18.89.
∫
0 ∞
− e − bx 1 ⎛ b 2 + p2 ⎞ dx = ln ⎜ 2 x sec px 2 ⎝ a + p2 ⎟⎠
∞ − ax
e
0
− e − bx b a dx = tan −1 − tan −1 x csc px p p
∞ − ax
e
0
∞ − ax
e
0
(1 − cos x ) a dx = cot −1 a − ln (a 2 + 1) 2 x2
DEFINITE INTEGRALS
114
Definite Integrals Involving Logarithmic Functions 18.90.
∫
1 0
x m (ln x )n dx =
(−1)n n ! (m + 1)n+1
m > −1,
n = 0, 1, 2,…
If n ≠ 0, 1, 2, … replace n! by Γ(n + 1). 18.91.
ln x π2 dx = − ∫0 1 + x 12
18.92.
ln x π2 dx = − ∫0 1 − x 6
18.93. 18.94. 18.95. 18.96. 18.97.
1
1
∫ ∫ ∫ ∫ ∫
ln (1 + x ) π2 dx = x 12
1 0
ln (1 − x ) π2 dx = − x 6
1 0 1 0 1 0
ln x ln (1 + x ) dx = 2 − 2 ln 2 − ln x ln (1 − x ) dx = 2 −
∞ 0
0 < p<1
xm − xn m +1 dx = ln ln x n +1
1
∫
0
18.99.
∫
0
18.100.
∫
18.101.
π2 ⎛ e x + 1⎞ ln dx = ∫0 ⎜⎝ e x − 1⎟⎠ 4
18.102.
∫
∞
e − x ln x dx = −γ
∞ 0
π (γ + 2 ln 2) 4
e − x ln x dx = − 2
∞
∫
π /2 0
π /2 0
π
18.104.
∫
0
18.105.
∫
0
18.106.
∫
0
18.107.
π2 6
x p−1 ln x dx = −π 2 csc pπ cot pπ 1+ x
18.98.
18.103.
π2 12
∫
2π
0
∫
π /2 0
(ln sin x )2 dx =
x ln sin x dx = −
π /2
π
ln sin x dx =
∫
ln cos x dx = − π /2
0
π ln 2 2
(ln cos x )2 dx =
π π3 (ln 2)2 + 2 24
π2 ln 2 2
sin x ln sin x dx = ln 2 − 1 ln (a + b sin x ) dx =
∫
2π 0
ln (a + b cos x ) dx = 2π ln (a + a 2 − b 2 )
⎛ a + a2 − b2 ⎞ ln(a + b cos x ) dx = π ln ⎜ ⎟ 2 ⎝ ⎠
DEFINITE INTEGRALS
π
18.108.
∫
18.109.
∫
0
18.110.
∫
0
18.111.
∫
0
⎧⎪2π ln a, a b > 0 ln (a 2 − 2ab cos x + b 2 ) dx = ⎨ ⎪⎩2π ln b, b a > 0
π /4
π /2
a 0
115
ln (1 + tan x ) dx =
π ln 2 8
⎛ 1 + b cos x ⎞ 1 sec x ln ⎜ dx = {(cos −1 a)2 − (cos −1 b)2} ⎟ 2 ⎝ 1 + a cos x ⎠
x⎞ ⎛ ⎛ sin a sin 2a sin 3a ⎞ ln ⎜ 2 sin ⎟ dx = − ⎜ 2 + + 2 + ⎟ 2⎠ 22 3 ⎝ ⎝ 1 ⎠
See also 18.102.
Definite Integrals Involving Hyperbolic Functions 18.112.
∫
18.113.
∫
18.114.
∫
18.115.
∫
∞ 0 ∞ 0 ∞ 0 ∞ 0
sin ax π aπ dx = tanh 2b 2b sinh bx cos ax π aπ dx = sech cosh bx 2b 2b x dx π2 = 2 sinh ax 4 a
{
}
x n dx 2n+1 − 1 1 1 1 = n n+1 Γ (n + 1) n+1 + n+1 + n+1 + sinh ax 2 a 1 2 3
If n is an odd positive integer, the series can be summed. 18.116.
∫
18.117.
∫
∞ 0 ∞ 0
1 sinh ax π aπ dx = csc − b 2a 2b e bx + 1 aπ sinh ax π 1 dx = cot − b 2a 2b e bx − 1
Miscellaneous Definite Integrals 18.118.
∫
f (ax ) − f (bx ) b dx = { f (0) − f (∞)}ln x a
∞ 0
This is called Frullani’s integral. It holds if f ′(x) is continuous and 18.119.
∫
18.120.
∫
1 0
dx 1 1 1 x = 1 + 2 + 3 + x 1 2 3
a −a
(a + x )m −1 (a − x )n−1 dx = (2a)m+ n−1
Γ (m ) Γ (n) Γ (m + n)
∫
∞ 0
f ( x ) − f (∞) dx converges. x
Section V: Differential Equations and Vector Analysis
19
BASIC DIFFERENTIAL EQUATIONS and SOLUTIONS
DIFFERENTIAL EQUATION
SOLUTION
19.1. Separation of variables f1 ( x ) g ( y) dx + ∫ 2 dy = c f2 ( x ) g1 ( y)
∫
f1(x) g1(y) dx + f2(x) g2(y) dy = 0
19.2. Linear first order equation ye ∫
dy + p( x ) y = Q( x ) dx
Pdx
= ∫ Qe ∫
Pdx
dx + c
19.3. Bernoulli’s equation dy + P( x ) y = Q( x ) y n dx
e
∫
(1− n ) Pdx
= (1 − n) ∫ Qe
∫
(1− n ) Pdx
dx + c
where y = y1−n. If n = 1, the solution is ln y = ∫ (Q − P) dx + c
19.4. Exact equation M(x, y)dx + N(x, y)dy = 0 where ∂M/∂y = ∂N/∂x.
⎛
∂
⎞
∫ M ∂x + ∫ ⎜⎝ N − ∂y ∫ M ∂x⎟⎠ dy = c where ∂x indicates that the integration is to be performed with respect to x keeping y constant.
19.5 Homogeneous equation
⎛ y⎞ dy = F⎜ ⎟ dx ⎝ x⎠
ln x =
d
∫ F ( ) − + c
where y = y/x. If F(y) = y, the solution is y = cx.
116 Copyright © 2009, 1999, 1968 by The McGraw-Hill Companies, Inc. Click here for terms of use.
BASIC DIFFERENTIAL EQUATIONS AND SOLUTIONS
117
19.6.
y F(xy) dx + x G(xy) dy = 0
ln x =
G ( ) d
∫ {G( ) − F ( )} + c
where y = xy. If G(y) = F(y), the solution is xy = c.
19.7. Linear, homogeneous second order equation
d2y dy +a + by = 0 2 dx dx
Let m1, m2 be the roots of m2 + am + b = 0. Then there are 3 cases. Case 1. m1, m2 real and distinct: y = c1e m x + c2 e m x 1
Case 2.
2
m1, m2 real and equal: y = c1 e m x + c2 x e m x 1
a, b are real constants. Case 3.
m1 = p + qi,
1
m2 = p − qi:
y = e px (c1 cos qx + c2 sin qx )
where p = −a/2, q = b − a 2 /4 . 19.8. Linear, nonhomogeneous second order equation
There are 3 cases corresponding to those of entry 19.7 above. Case 1. y = c1e m x + c2 e m x
2
d y dy +a + by = R( x ) 2 dx dx
1
+
a, b are real constants.
2
em x e − m x R( x ) dx m1 − m2 ∫ 1
1
+
em x e − m x R( x ) dx m2 − m1 ∫ 2
2
Case 2. y = c1 e m x + c2 x e m x 1
1
+ xe m x ∫ e − m x R( x ) dx 1
1
− e m x ∫ xe − m x R( x ) dx 1
1
Case 3. y = e px (c1 cos qx + c2 sin qx ) +
e px sin qx − px ∫ e R( x ) cos qx dx q −
e px cos qx − px ∫ e R( x ) sin qx dx q
BASIC DIFFERENTIAL EQUATIONS AND SOLUTIONS
118
19.9. Euler or Cauchy equation x2
Putting x = et, the equation becomes d2y dy + (a − 1) + by = S (e t ) dt 2 dt
d2y dy + ax + by = S ( x ) 2 dx dx
and can then be solved as in entries 19.7 and 19.8 above. 19.10. Bessel’s equation x2
d2y dy +x + ( 2 x 2 − n 2 ) y = 0 2 dx dx
y = c1 J n ( x ) + c2 Yn ( x )
See 27.1 to 27.15.
19.11. Transformed Bessel’s equation
x2
d2y dy + (2 p + 1) x + (a 2 x 2 r + β 2 ) y = 0 dx 2 dx
⎛α ⎞ ⎛ α ⎞ ⎪⎫ ⎪⎧ y = x − p ⎨c1 J q/r ⎜ x r ⎟ + c2 Yq/r ⎜ x r ⎟ ⎬ ⎝r ⎠ ⎝ r ⎠ ⎪⎭ ⎪⎩
where q =
p2 − β 2 .
19.12. Legendre’s equation (1 − x 2 )
d2y dy − 2x + n(n + 1) y = 0 dx 2 dx
y = c1 Pn ( x ) + c2 Qn ( x )
See 28.1 to 28.48.
20
FORMULAS from VECTOR ANALYSIS
Vectors and Scalars Various quantities in physics such as temperature, volume, and speed can be specified by a real number. Such quantities are called scalars. Other quantities such as force, velocity, and momentum require for their specification a direction as well as magnitude. Such quantities are called vectors. A vector is represented by an arrow or directed line segment indicating direction. The magnitude of the vector is determined by the length of the arrow, using an appropriate unit.
Notation for Vectors A vector is denoted by a bold faced letter such as A (Fig. 20.1). The magnitude is denoted by |A| or A. The tail end of the arrow is called the initial point, while the head is called the terminal point.
Fundamental Definitions 1.
Equality of vectors. Two vectors are equal if they have the same magnitude and direction. Thus, A = B in (Fig. 20-1).
2.
Multiplication of a vector by a scalar. If m is any real number (scalar), then mA is a vector whose magnitude is |m| times the magnitude of A and whose direction is the same as or opposite to A according as m > 0 or m < 0. If m = 0, then mA = 0 is called the zero or null vector.
Fig. 20-1
3. Sums of vectors. The sum or resultant of A and B is a vector C = A + B formed by placing the initial point B on the terminal point A and joining the initial point of A to the terminal point of B as in Fig. 20-2b. This definition is equivalent to the parallelogram law for vector addition as indicated in Fig. 20-2c. The vector A − B is defined as A + (−B).
Fig. 20-2
Extension to sums of more than two vectors are immediate. Thus, Fig. 20-3 shows how to obtain the sum E of the vectors A, B, C, and D.
119
FORMULAS FROM VECTOR ANALYSIS
120
Fig. 20-3
4.
Unit vectors. A unit vector is a vector with unit magnitude. If A is a vector, then a unit vector in the direction of A is a = A/A where A > 0.
Laws of Vector Algebra If A, B, C are vectors and m, n are scalars, then: 20.1. A + B = B + A
Commutative law for addition
20.2. A + (B + C) = (A + B) + C
Associative law for addition
20.3. m(nA) = (mn)A = n(mA)
Associative law for scalar multiplication
20.4. (m + n)A = mA + nA
Distributive law
20.5. m(A + B) = mA + mB
Distributive law
Components of a Vector A vector A can be represented with initial point at the origin of a rectangular coordinate system. If i, j, k are unit vectors in the directions of the positive x, y, z axes, then
z A
k
20.6. A = A1i + A2j + A3k
i
where A1i, A2j, A3k are called component vectors of A in the i, j, k directions and A1, A2, A3 are called the components of A.
Dot or Scalar Product 20.7. A • B = AB cos q
0qp
where q is the angle between A and B. Fundamental results follow:
j
A3k A2j
x
Fig. 20-4
y A2i
FORMULAS FROM VECTOR ANALYSIS
20.8. A • B = B • A
Commutative law
20.9. A • (B + C) = A • B + A • C
Distributive law
121
20.10. A • B = A1B1 + A2B2 + A3B3 where A = A1i + A2j + A3k, B = B1i + B2j + B3k.
Cross or Vector Product 20.11. A × B = AB sin q u
0qp
where q is the angle between A and B and u is a unit vector perpendicular to the plane of A and B such that A, B, u form a right-handed system (i.e., a right-threaded screw rotated through an angle less than 180° from A to B will advance in the direction of u as in Fig. 20-5). Fundamental results follow: i A × = A B 20.12. 1 B1
j A2 B2
k A3 B3
u
B
A
= ( A2 B3 − A3 B2 )i + ( A3 B1 − A1 B3 ) j + ( A1 B2 − A2 B1 )k
Fig. 20-5
20.13. A × B = − (B × A) 20.14. A × (B + C) = A × B + A × C 20.15. | A × B | = area of parallelogram having sides A and B
Miscellaneous Formulas Involving Dot and Cross Products 20.16.
A1 A2 A i (B × C) = B1 B2 C1 C2
A3 B3 = A1 B2C3 + A2 B3C1 + A3 B1C2 − A3 B2C1 − A2 B1C3 − A1 B3C2 C3
20.17. | A • (B × C) | = volume of parallelepiped with sides A, B, C 20.18. A × (B × C) = B(A • C) − C(A • B) 20.19. (A × B) × C = B(A • C) − A(B • C) 20.20.
(A × B) • (C × D) = (A • C)(B • D) − (A • D)(B • C)
20.21. (A × B) × (C × D) = C{A • (B × D)}− D{A • (B × C)} = B{A • (C × D)} − A{B • (C × D)}
FORMULAS FROM VECTOR ANALYSIS
122
Derivatives of Vectors The derivative of a vector function A(u) = A1(u)i + A2(u)j + A3(u)k of the scalar variable u is given by 20.22.
dA dA dA A(u + Δu) − A(u) dA1 = lim = i+ 2 j+ 3 k Δ u → 0 du du Δu du du
Partial derivatives of a vector function A(x, y, z) are similarly defined. We assume that all derivatives exist unless otherwise specified.
Formulas Involving Derivatives 20.23.
d dB dA (A i B) = A i + iB du du du
20.24.
d dB dA (A × B) = A × + ×B du du du
20.25.
⎛ dB ⎛ ⎞ d dA dC ⎞ {A i (B × C )} = i (B × C ) + A i ⎜ × C⎟ + A i ⎜ B × du du du ⎟⎠ ⎝ du ⎝ ⎠
20.26.
Ai
dA dA =A du du
20.27.
Ai
dA =0 du
if | A | is a constant
The Del Operator The operator del is defined by
20.28.
∇=i
∂ ∂ ∂ + j +k ∂x ∂y ∂z
In the following results we assume that U = U(x, y, z), V = V(x, y, z), A = A(x, y, z) and B = B(x, y, z) have partial derivatives.
The Gradient ⎛ ∂ ∂ ∂⎞ ∂U ∂U ∂U 20.29. Gradient of U = grad U = ∇U = ⎜ i + j + k ⎟ U = i+ j+ k ∂y ∂z ⎠ ∂x ∂y ∂z ⎝ ∂x
The Divergence ⎛ ∂ ∂ ∂⎞ 20.30. Divergence of A = div A = ∇ • A = ⎜ i + j + k ⎟ i ( A1 i + A2 j + A3 k) ∂y ∂z ⎠ ⎝ ∂x =
∂A1 ∂A2 ∂A3 + + ∂y ∂z ∂x
FORMULAS FROM VECTOR ANALYSIS
The Curl 20.31. Curl of A = curl A = ∇ × A ⎛ ∂ ∂ ∂⎞ = ⎜ i + j + k ⎟ × ( A1 i + A2 j + A3 k) ∂ ∂ ∂ x y z⎠ ⎝ i ∂ = ∂x A1
j ∂ ∂y A2
k ∂ ∂z A3
⎛ ∂A ∂A ⎞ ⎛ ∂A ∂A ⎞ ⎛ ∂A ∂A ⎞ = ⎜ 3 − 2 ⎟ i + ⎜ 1 − 3⎟ j + ⎜ 2 − 1⎟ k ∂x ⎠ ∂y ⎠ ∂z ⎠ ⎝ ∂x ⎝ ∂z ⎝ ∂y
The Laplacian 20.32. Laplacian of U = ∇ 2U = ∇ i (∇U ) = 20.33.
Laplacian of A = ∇ 2 A =
∂ 2U ∂ 2U ∂ 2U + + 2 ∂x 2 ∂y 2 ∂z
∂2 A ∂2 A ∂2 A + 2 + 2 ∂x 2 ∂y ∂z
The Biharmonic Operator 20.34. Biharmonic operator on U = ∇ 4U = ∇ 2 (∇ 2U ) =
∂ 4U ∂ 4U ∂ 4U ∂ 4U ∂ 4U ∂ 4U + + + 2 + 2 2 + ∂x 4 ∂y 4 ∂z 4 ∂x 2 ∂y 2 ∂y 2 ∂z 2 ∂x 2 ∂z 2
Miscellaneous Formulas Involving ∇ 20.35.
∇(U + V ) = ∇U + ∇V
20.36.
∇ i (A + B) = ∇ i A + ∇ i B
20.37.
∇ × (A + B) = ∇ × A + ∇ × B
20.38.
∇ i (UA) = (∇U ) i A + U (∇ i A)
20.39.
∇ × (UA) = (∇U ) × A + U (∇ × A)
20.40.
∇ i (A × B) = B i (∇ × A) − A i (∇ × B)
20.41.
∇ × (A × B) = (B i ∇) A − B(∇ i A) − (A i ∇)B + A(∇ i B)
20.42.
∇(A i B) = (B i ∇) A + ( A i ∇)B + B × (∇ × A) + A × (∇ × B)
20.43.
∇ × (∇U ) = 0,
20.44.
∇ i (∇ × A) = 0, that is, the divergence of the curl of A is zero.
20.45.
∇ × (∇ × A) = ∇(∇ i A) − ∇ 2 A
that is, the curl of the gradient of U is zero.
123
FORMULAS FROM VECTOR ANALYSIS
124
Integrals Involving Vectors d B(u). then the indefinite integral of A(u) is as follows: du
If A(u) = 20.46.
∫ A(u)du = B(u) + c,
c = constant vector
The definite integral of A(u) from u = a to u = b in this case is given by 20.47.
∫
b
A(u) du = B(b) − B(a)
a
The definite integral can be defined as in 18.1.
Line Integrals z
Consider a space curve C joining two points P1(a1, a2, a3) and P2(b1, b2, b3) as in Fig. 20-6. Divide the curve into n parts by points of subdivision (x1, y1, z1), . . . , (xn−1, yn−1, zn−1). Then the line integral of a vector A(x, y, z) along C is defined as
P2 C (xp , yp , zp)
P1
20.48.
∫
c
A i dr =
∫
P2
n→∞
P1
y
n
A i dr = lim ∑ A( x p , y p , z p ) i Δrp p =1
x
Fig. 20-6
where Δrp = Δx p i + Δy p j + Δz pk, Δx p = x p+1 − x p , Δy p = y p+1 − y p , Δz p = z p+1 − z p and where it is assumed that as n → ∞ the largest of the magnitudes |rp | approaches zero. The result 20.48 is a generalization of the ordinary definite integral (see 18.1). The line integral 20.48 can also be written as 20.49.
∫
C
A i dr =
∫
( A1dx + A2 dy + A3 dz )
C
using A = A1i + A2j + A3k
dr = dxi + dyj + dzk.
and
Properties of Line Integrals 20.50. 20.51.
p2
∫
p1
∫
P1
P1
A i dr = − ∫ A i dr P2
P2
A i dr =
∫
P3 P1
P2
A i dr + ∫ A i dr P3
Independence of the Path In general, a line integral has a value that depends on the particular path C joining points P1 and P2 in a region . However, in the case of A = ∇f or ∇ × A = 0 where f and its partial derivatives are continuous in , the line integral ∫ A i dr is independent of the path. In such a case, C
20.52.
∫
C
A i dr =
∫
P2 P1
A i dr = φ (P2 ) − φ (P1 )
where f(P1) and f(P2) denote the values of f at P1 and P2, respectively. In particular if C is a closed curve,
FORMULAS FROM VECTOR ANALYSIS
20.53.
∫
C
125
A i dr = ∫ A i dr = 0 C
where the circle on the integral sign is used to emphasize that C is closed.
Multiple Integrals Let F(x, y) be a function defined in a region of the xy plane as in Fig. 20-7. Subdivide the region into n parts by lines parallel to the x and y axes as indicated. Let Ap = xp yp denote an area of one of these parts. Then the integral of F(x, y) over is defined as 20.54.
∫
n
F( x , y) dA = lim ∑ F( x p , y p )ΔAp n→∞
p =1
provided this limit exists. In such a case, the integral can also be written as 20.55.
b
∫ ∫ x =a
f2 (x) y = f1 (x)
Fig. 20-7
F( x , y) dy dx
{∫
⎫ F( x , y) dy⎬ dx ⎭ where y = f1(x) and y = f2(x) are the equations of curves PHQ and PGQ, respectively, and a and b are the x coordinates of points P and Q. The result can also be written as =
20.56.
∫
b
x =a
d
g2( y )
y =c
x = g1( y )
∫ ∫
f2( x )
y = f1( x )
F ( x , y) dx dy =
∫
d y =c
{∫
g2( y ) x = g1( y )
}
F ( x , y) dx dy
where x = g1(y), x = g2(y) are the equations of curves HPG and HQG, respectively, and c and d are the y coordinates of H and G. These are called double integrals or area integrals. The ideas can be similarly extended to triple or volume integrals or to higher multiple integrals.
Surface Integrals z
Subdivide the surface S (see Fig. 20-8) into n elements of area ΔS p , p = 1, 2, … , n, Let A( x p , y p , z p ) = A p where ( x p , y p , z p ) is a point P in Sp. Let Np be a unit normal to Sp at P. Then the surface integral of the normal component of A over S is defined as 20.57.
∫
Np
γ
ΔSp S
n
S
A i N dS = lim ∑ A p i N p ΔS p n→∞
y
p =1
Δ xp Δyp x
Fig. 20-8
FORMULAS FROM VECTOR ANALYSIS
126
Relation Between Surface and Double Integrals If is the projection of S on the xy plane, then (see Fig. 20-8) 20.58.
∫
S
A i N dS = ∫
∫ AiN
dx dy Nik
The Divergence Theorem Let S be a closed surface bounding a region of volume V; and suppose N is the positive (outward drawn) normal and dS = N dS. Then (see Fig. 20-9) 20.59.
∫
V
∇ i A dV =
∫
S
A i dS
The result is also called Gauss’ theorem or Green’s theorem.
z
z N
N S
dS S
dS
C
y
y
x
x
Fig. 20-10
Fig. 20-9
Stokes’ Theorem Let S be an open two-sided surface bounded by a closed non-intersecting curve C(simple closed curve) as in Fig. 20-10. Then 20.60.
∫
C
A i dr =
∫
S
(∇ × A) i dS
where the circle on the integral is used to emphasize that C is closed.
Green’s Theorem in the Plane 20.61.
∫
C
(P dx + Q dy) =
∫
⎛ ∂Q ∂P ⎞ − dx dy R⎜ ⎝ ∂x ∂y ⎟⎠
where R is the area bounded by the closed curve C. This result is a special case of the divergence theorem or Stokes’ theorem.
FORMULAS FROM VECTOR ANALYSIS
127
Green’s First Identity 20.62.
∫
V
{(φ ∇ 2ψ + ( ∇φ ) i (∇ψ )}dV = ∫ (φ ∇ψ ) i dS
where f and y are scalar functions.
Green’s Second Identity 20.63.
∫
V
(φ ∇ 2ψ − ψ ∇ 2φ ) dV =
∫
S
(φ ∇ψ − ψ ∇φ ) i dS
Miscellaneous Integral Theorems 20.64.
∫
∇ × A dV =
∫
φ dr =
V
20.65.
C
∫
S
∫
S
dS × A
dS × ∇φ
Curvilinear Coordinates A point P in space (see Fig. 20-11) can be located by rectangular coordinates (x, y, z,) or curvilinear coordinates (u1, u2, u3) where the transformation equations from one set of coordinates to the other are given by 20.66.
z u3 curve e3
x = x (u1, u2 , u3)
c2 u2 =
y = y(u1, u2 , u3) z = z (u1, u2 , u3)
u1 curve
e1
P
u3 = c3
u1 = c1
e2 u2 curve y
If u2 and u3 are constant, then as u1 varies, the position vector r = xi + yj + zk of P describes a curve called the u1 coordinate curve. Similarly, we define the u2 and u3 coordinate curves through P. The vectors ∂r/∂u1, ∂r/∂u2 , ∂r/∂u3 represent tangent vectors to the u1, u2, u3 coordinate curves. Letting e1, e2, e3 be unit tangent vectors to these curves, we have 20.67.
∂r = h1e1 , ∂u1
∂r = h2 e 2 , ∂u2
x
Fig. 20-11
∂r = h3 e3 ∂u3
where 20.68.
h1 =
∂r , ∂u1
h2 =
∂r , ∂u2
h3 =
∂r ∂u3
are called scale factors. If e1, e2, e3 are mutually perpendicular, the curvilinear coordinate system is called orthogonal.
FORMULAS FROM VECTOR ANALYSIS
128
Formulas Involving Orthogonal Curvilinear Coordinates ∂r ∂r ∂r du + du + du = h1 du1e1 + h2 du2 e 2 + h3 du3 e 3 ∂u1 1 ∂u2 2 ∂u3 3
20.69.
dr =
20.70.
ds 2 = dr i dr = h12 du12 + h22 du22 + h32 du32
where ds is the element of are length. If dV is the element of volume, then 20.71.
dV = | (h1e1du1 ) i (h2 e 2 du2 ) × (h3 e 3 du3 ) | = h1h2 h3 du1du2 du3 =
∂r ∂r ∂r ∂( x , y, z ) du1 du2 du3 = du du du × i ∂(u1, u2 , u3 ) 1 2 3 ∂u1 ∂u2 ∂u3
where 20.72.
∂x /∂u1 ∂( x , y, z ) = ∂y/∂u1 ∂(u1, u2 , u3 ) ∂z/∂u1
∂x /∂u2 ∂y/∂u2 ∂z/∂u2
∂x /∂u3 ∂y/∂u3 ∂z/du3
sometimes written J(x, y, z; u1, u2, u3), is called the Jacobian of the transformation.
Transformation of Multiple Integrals Result 20.72 can be used to transform multiple integrals from rectangular to curvilinear coordinates. For example, we have 20.73.
∫ ∫ ∫ F ( x , y, z) dx dy dz = ∫ ∫ ∫ G(u , u , u ) 1
2
3
′
∂( x , y, z ) du du du ∂(u1, u2 , u3) 1 2 3
where ′ is the region into which is mapped by the transformation and G(u1, u2, u3) is the value of F(x, y, z) corresponding to the transformation.
Gradient, Divergence, Curl, and Laplacian In the following, Φ is a scalar function and A = A1e1 + A2e2 + A3e3 is a vector function of orthogonal curvilinear coordinates u1, u2, u3. 20.74. Gradient of Φ = grad Φ = ∇Φ =
e1 ∂ Φ e 2 ∂ Φ e 3 ∂ Φ + + h1 ∂u1 h2 du2 h3 ∂u3 1 h1h2 h3
20.75.
Divergence of A = div A = ∇ i A =
20.76.
h1e1 1 ∂ Curl of A = curl A = ∇ × A = h1h2 h3 ∂u1 h1 A1 =
1 h2 h3
⎤ ⎡ ∂ ∂ ∂ (h1h2 A3 )⎥ (h2 h3 A1 ) + (h3 h1 A2 ) + ⎢ ∂u3 ∂u2 ⎦ ⎣ ∂u1 h2 e 2 ∂ ∂u2 h2 A2
h3 e 3 ∂ ∂u3 h3 A3
⎤ ⎤ ⎡ ∂ 1 ⎡ ∂ ∂ ∂ (h3 A3 ) − (h2 A2 )⎥ e1 + (h1 A1 ) − (h3 A3 )⎥ e 2 ⎢ ⎢ ∂ u u ∂ h h u u ∂ ∂ 1 3 ⎣ 3 1 3 ⎦ ⎦ ⎣ 2 +
1 h1h2
⎤ ⎡ ∂ ∂ (h2 A2 ) − (h1 A1 )⎥ e 3 ⎢ ∂u2 ⎦ ⎣ ∂u1
FORMULAS FROM VECTOR ANALYSIS
20.77. Laplacian of Φ = ∇ 2 Φ =
1 h1h2 h3
129
⎡ ∂ ⎛ h2 h3 ∂ Φ⎞ ∂ ⎛ h1h2 ∂ Φ⎞ ⎤ ∂ ⎛ h3 h1 ∂ Φ⎞ + + ⎢ ⎟⎥ ⎟ ⎜ ⎜ ⎟ ⎜ ⎢⎣ ∂u1 ⎝ h1 ∂u1 ⎠ ∂u2 ⎝ h2 ∂u2 ⎠ ∂u3 ⎝ h3 ∂u3 ⎠ ⎥⎦
Note that the biharmonic operator ∇ 4 Φ = ∇ 2 (∇ 2 Φ) can be obtained from 20.77.
Special Orthogonal Coordinate Systems Cylindrical Coordinates (r, q, z) (See Fig. 20-12) 20.78. x = r cos q, 20.79.
h12 = 1,
20.80.
∇2Φ =
y = r sin q,
h22 = r 2 ,
z=z
h32 = 1
∂2 Φ 1 ∂ Φ 1 ∂2 Φ ∂2 Φ + + + 2 ∂ r 2 r ∂r r 2 ∂θ 2 ∂z ez
z
z
er
eq
ef
(r, q, z)
P
P(r, q, f,)
er
z
z
x
x q y
y
r
f y
eq
x
x
Fig. 20-12.
Fig. 20-13.
Cylindrical coordinates.
Spherical coordinates.
Spherical Coordinates (r, q, f) (See Fig. 20-13)
20.81. x = r sin q cos f,
y = r sin q sin f,
20.82.
h12 = 1,
h32 = r 2 sin 2 θ
20.83.
∇2Φ =
h22 = r 2 ,
z = r cos q
1 ∂ ⎛ 2 ∂Φ ⎞ 1 1 ∂2 Φ ∂ ⎛ ∂Φ ⎞ + 2 2 + 2 r sin θ ⎜ ⎟ ⎜ ⎟ 2 r ∂r ⎝ dr ⎠ r sin θ ∂θ ⎝ ∂θ ⎠ r sin θ ∂φ 2
Parabolic Cylindrical Coordinates (u, y, z) 20.84.
x = 12 (u 2 − 2 ),
20.85.
h12 = h22 = u 2 + 2 ,
20.86.
∇2Φ =
y = u ,
z=z
h32 = 1
1 ⎛ ∂2 Φ ∂2 Φ ⎞ ∂2 Φ + + u 2 + 2 ⎜⎝ ∂u 2 ∂ 2 ⎟⎠ ∂z 2
The traces of the coordinate surfaces on the xy plane are shown in Fig. 20-14. They are confocal parabolas with a common axis.
Fig. 20-14
y
FORMULAS FROM VECTOR ANALYSIS
130
Paraboloidal Coordinates (u, y, f) 20.87.
x = u cos φ ,
where
y = u sin φ ,
u 0,
20.88.
h12 = h22 = u 2 + 2 ,
20.89.
∇2Φ =
0,
z = 12 (u 2 − 2 )
0 φ < 2π
h32 = u 2 2
∂ ⎛ ∂Φ ⎞ 1 1 1 ∂2 Φ ∂ ⎛ ∂Φ ⎞ u + + u(u 2 + 2 ) ∂u ⎜⎝ ∂u ⎟⎠ (u 2 + 2 ) ∂ ⎜⎝ ∂ ⎟⎠ u 2 2 ∂φ 2
Two sets of coordinate surfaces are obtained by revolving the parabolas of Fig. 20-14 about the x axis which is then relabeled the z axis. Elliptic Cylindrical Coordinates (u, y, z) 20.90.
x = a cosh u cos ,
where
u 0,
y = a sinh u sin , 0 < 2π ,
20.91.
h12 = h22 = a 2 (sinh 2 u + sin 2 ),
20.92.
∇2Φ =
z=z
− ∞< z < ∞ h32 = 1
⎛ ∂2 Φ ∂2 Φ ⎞ ∂2 Φ 1 + + 2 a (sinh u + sin ) ⎜⎝ ∂u 2 ∂ 2 ⎟⎠ ∂z 2 2
2
The traces of the coordinate surfaces on the xy plane are shown in Fig. 20-15. They are confocal ellipses and hyperbolas.
Fig. 20-15.
Elliptic cylindrical coordinates.
FORMULAS FROM VECTOR ANALYSIS
131
Prolate Spheroidal Coordinates (x, h, f) 20.93.
x = a sinh ξ sin η cos φ ,
y = a sinh ξ sin η sin φ ,
ξ 0,
where
20.94.
h12 = h22 = a 2 (sinh 2 ξ sin 2 η),
20.95.
∇2Φ =
z = a cosh ξ cos η
0 η π,
0 φ < 2π
h32 = a 2 sinh 2 ξ sin 2 η
∂ ⎛ ∂Φ ⎞ 1 sinh ξ ∂ξ ⎟⎠ a 2 (sinh 2 ξ + sin 2 η) sinh ξ ∂ξ ⎜⎝ +
∂2 Φ ∂ ⎛ ∂Φ ⎞ 1 1 sin η + ⎜ ⎟ a 2 (sinh 2 ξ + sin 2 η) sin η ∂η ⎝ ∂η ⎠ a 2 sinh 2 ξ sin 2 η ∂φ 2
Two sets of coordinate surfaces are obtained by revolving the curves of Fig. 20-15 about the x axis which is relabeled the z axis. The third set of coordinate surfaces consists of planes passing through this axis. Oblate Spheroidal Coordinates (x, h, f) 20.96.
x = a cosh ξ cos η cos φ ,
y = a cosh ξ cos η sin φ ,
where
ξ 0,
20.97.
h12 = h22 = a 2 (sinh 2 ξ + sin 2 η),
20.98.
∇2Φ =
z = a sinh ξ sin η
− π /2 η π /2,
0 φ < 2π
h32 = a 2 cosh 2 ξ cos 2 η
∂ ⎛ ∂Φ ⎞ 1 cosh ξ ⎜ 2 ∂ξ ⎠⎟ a (sinh ξ + sin η) cosh ξ ∂ξ ⎝ 2
2
+
∂ ⎛ ∂Φ ⎞ ∂2 Φ 1 1 cos η + 2 ⎜ ⎟ 2 2 2 ∂η ⎠ a cosh ξ cos η ∂φ 2 a (sinh ξ + sin η) cos η ∂η ⎝ 2
2
Two sets of coordinate surfaces are obtained by revolving the curves of Fig. 20-15 about the y axis which is relabeled the z axis. The third set of coordinate surfaces are planes passing through this axis. Bipolar Coordinates (u, y, z)
20.99.
x=
a sinh , cosh − cos u
where
y=
a sin u , cosh − cos u 0 u < 2π ,
z=z −∞< < ∞,
−∞< z < ∞
or 20.100.
x 2 + (y − a cot u)2 = a 2 csc 2 u,
( x − a coth )2 + y 2 = a 2 csch 2 ,
z=z
FORMULAS FROM VECTOR ANALYSIS
132
20.101.
h12 = h22 =
20.102.
∇2Φ =
a2 , (cosh − cos u)2
(cosh − cos u)2 a2
h32 = 1
⎛ ∂2 Φ ∂2 Φ ⎞ ∂2 Φ ⎜⎝ ∂u 2 + ∂ 2 ⎟⎠ + ∂z 2
The traces of the coordinate surfaces on the xy plane are shown in Fig. 20-16.
Fig. 20-16.
Bipolar coordinates.
Toroidal Coordinates (u, y, f) a sinh cos φ , cosh − cos u
20.103.
x=
20.104.
h12 = h22 =
20.105.
∇ 2Φ =
y=
a sinh sin φ , cosh − cos u
a2 , (cosh − cos u)2
h32 =
z=
a sin u cosh − cos u
a 2 sinh 2 (cosh − cos u)2
(cosh − cos u)3 ∂ ⎛ 1 ∂Φ⎞ 2 ⎜ a ∂u ⎝ cosh − cos u ∂u ⎟⎠ +
(cosh − cos u)3 ∂ ⎛ sinh ∂Φ⎞ (cosh − cos u)2 ∂ 2Φ + a 2 sinh ∂ ⎜⎝ cosh − cos u ∂ ⎟⎠ a 2 sinh 2 ∂f 2
The coordinate surfaces are obtained by revolving the curves of Fig. 20.16 about the y axis which is relabeled the z axis. Conical Coordinates (l, m, v)
μ v , ab
20.106.
x=
20.107.
h12 = 1,
y= h22 =
(μ 2 − a 2 )(v 2 − a 2 ) , a a2 − b2
2 (μ 2 − v 2 ) , (μ 2 − a 2 )(b 2 − μ 2 )
h32 =
z=
(μ 2 − b 2 )(v 2 − b 2 ) b b2 − a2
2 (μ 2 − v 2 ) (v 2 − a 2 )(v 2 − b 2 )
FORMULAS FROM VECTOR ANALYSIS
133
Confocal Ellipsoidal Coordinates (l, m, y) ⎧ x2 y2 z2 ⎪ a 2 − + b 2 − + c 2 − = 1, ⎪ z2 y2 ⎪ x2 + 2 + 2 = 1, ⎨ 2 ⎪a − μ b − μ c − μ ⎪ x2 y2 z2 = 1, + 2 + 2 ⎪ 2 ⎩a − v b − v c − v
20.108.
< c2 < b2 < a2 c2 < μ < b2 < a2 c2 < b2 < v < a2
or
20.109.
⎧ 2 (a 2 − )(a 2 − μ )(a 2 − v) ⎪x = (a 2 − b 2 )(a 2 − c 2 ) ⎪ ⎪ 2 (b 2 − )(b 2 − μ )(b 2 − v) ⎨y = (b 2 − a 2 )(a 2 − c 2 ) ⎪ ⎪ (c 2 − )(c 2 − μ )(c 2 − v) ⎪z 2 = (c 2 − a 2 )(c 2 − b 2 ) ⎩
20.110.
⎧ 2 ⎪h1 = 4(a 2 ⎪ ⎪ 2 ⎨h2 = 4(a 2 ⎪ ⎪ ⎪h32 = 4(a 2 ⎩
(μ − )(v − ) − )(b 2 − )(c 2 − ) (v − μ )( − μ ) − μ )(b 2 − μ )(c 2 − μ ) ( − v)(μ − v) − v)(b 2 − v)(c 2 − v)
Confocal Paraboloidal Coordinates (l, m, v) ⎧ x2 h2 ⎪ a 2 − + b 2 − = z − , ⎪ y2 ⎪ x2 + 2 = z − μ, ⎨ 2 ⎪a − μ b − μ ⎪ x2 y2 + 2 = z − v, ⎪ 2 ⎩a − v b − v or
20.111.
20.112.
.
20.113.
− ∞< < b 2 b2 < μ < a2 a2 < v < ∞
⎧ 2 (a 2 − )(a 2 − μ )(a 2 − v) ⎪x = b2 − a2 ⎪⎪ ⎨ 2 (b 2 − )(b 2 − μ )(b 2 − v) ⎪y = a2 − b2 ⎪ ⎪⎩z = + μ + v − a 2 − b 2 ⎧ 2 ( μ − )(v − ) ⎪h1 = 4(a 2 − )(b 2 − ) ⎪ ⎪ 2 (v − μ )( − μ ) ⎨h2 = 4 a 2 − μ )(b 2 − μ ) ( ⎪ ⎪ ( − v)(μ − v) ⎪h32 = 16(a 2 − v)(b 2 − v) ⎩
Section VI: Series
21
SERIES of CONSTANTS
Arithmetic Series 21.1.
a + (a + d ) + (a + 2 d ) + ⋅⋅⋅ + {a + (n − 1)d } = 12 n{2 a + (n − 1)d } = 12 n(a + l )
where l a (n 1)d is the last term. Some special cases are 21.2.
1 + 2 + 3 + ⋅⋅⋅ + n = 12 n(n + 1)
21.3.
1 + 3 + 5 + ⋅⋅⋅ + (2 n − 1) = n 2
Geometric Series 21.4.
a + ar + ar 2 + ar 3 + ⋅⋅⋅ + ar n−1 =
a(1 − r n ) a − rl = 1− r 1− r
where l ar n1 is the last term and r ≠ 1. If 1 < r < 1, then 21.5.
a + ar + ar 2 + ar 3 + ⋅⋅⋅ =
a 1− r
Arithmetic-Geometric Series 21.6.
a + (a + d )r + (a + 2 d )r 2 + ⋅⋅⋅ + {a + (n − 1)d }r n−1 =
a(1 − r n ) rd {1 − nr n−1 + (n − 1)r n } + 1− r (1 − r )2
where r ≠ 1. If 1 < r < 1, then 21.7.
a + (a + d )r + (d + 2 d )r 2 + ⋅⋅⋅ =
a rd + 1 − r (1 − r )2
Sums of Powers of Positive Integers 21.8.
1p + 2 p + 3 p + ⋅⋅⋅ + n p =
n p+1 1 p B1 pn p−1 B2 p( p − 1)( p − 2 )n p− 3 + ⋅⋅⋅ + n + − 4! p +1 2 2!
where the series terminates at n2 or n according as p is odd or even, and Bk are the Bernoulli numbers (see page 142).
134 Copyright © 2009, 1999, 1968 by The McGraw-Hill Companies, Inc. Click here for terms of use.
SERIES OF CONSTANTS
135
Some special cases are 21.9.
1 + 2 + 3 + ⋅⋅⋅ + n =
n(n + 1) 2
21.10.
12 + 2 2 + 32 + ⋅⋅⋅ + n 2 =
n(n + 1)(2 n + 1) 6
21.11.
13 + 2 3 + 33 + ⋅⋅⋅ + n 3 =
n 2 (n + 1)2 = (1 + 2 + 3 + ⋅⋅⋅ + n )2 4
21.12.
14 + 2 4 + 34 + ⋅⋅⋅ + n 4 =
n(n + 1)(2 n + 1)(3n 2 + 3n − 1) 30
If Sk = 1k + 2 k + 3k + ⋅⋅⋅ + n k where k and n are positive integers, then 21.13.
⎛ k + 1⎞ ⎛ k + 1⎞ ⎛ k + 1⎞ k +1 ⎜⎝ 1 ⎟⎠ S1 + ⎜⎝ 2 ⎟⎠ S2 + ⋅⋅⋅ + ⎜⎝ k ⎟⎠ Sk = (n + 1) − (n + 1)
Series Involving Reciprocals of Powers of Positive Integers 21.14.
1−
1 1 1 1 + − + − ⋅⋅⋅ = ln 2 2 3 4 5
21.15.
1−
π 1 1 1 1 + − + − ⋅⋅⋅ = 3 5 7 9 4
21.16.
1−
π 3 1 1 1 1 1 + − + − ⋅⋅⋅ = + ln 2 4 7 10 13 9 3
21.17.
1−
π 2 1 1 1 1 + − + − ⋅⋅⋅ = + 5 9 13 17 8
21.18.
1 1 1 1 1 π 3 1 − + − + − ⋅⋅⋅ = + ln 2 2 5 8 11 14 9 3
21.19.
1 1 1 1 π2 2 + 2 + 2 + 2 + ⋅⋅⋅ = 6 1 2 3 4
21.20.
1 1 1 1 π4 4 + 4 + 4 + 4 + ⋅⋅⋅ = 90 1 2 3 4
21.21.
1 1 1 1 π6 6 + 6 + 6 + 6 + ⋅⋅⋅ = 945 1 2 3 4
21.22.
1 1 1 1 π2 2 − 2 + 2 − 2 + ⋅⋅⋅ = 12 1 2 3 4
21.23.
1 1 1 1 7π 4 4 − 4 + 4 − 4 + ⋅⋅⋅ = 720 1 2 3 4
21.24.
1 1 1 1 31π 6 6 − 6 + 6 − 6 + ⋅⋅⋅ = 30, 240 1 2 3 4
21.25.
1 1 1 1 π2 2 + 2 + 2 + 2 + ⋅⋅⋅ = 8 1 3 5 7
21.26.
1 1 1 1 π4 4 + 4 + 4 + 4 + ⋅⋅⋅ = 96 1 3 5 7
2 ln (1 + 2 ) 4
SERIES OF CONSTANTS
136
21.27.
1 1 1 1 π6 6 + 6 + 6 + 6 + ⋅⋅⋅ = 960 1 3 5 7
21.28.
1 1 1 1 π3 3 − 3 + 3 − 3 + ⋅⋅⋅ = 32 1 3 5 7
21.29.
1 1 1 1 3π 3 2 3 + 3 − 3 − 3 + ⋅⋅⋅ = 128 1 3 5 7
21.30.
1 1 1 1 1 + + + + ⋅⋅⋅ = 1i 3 3 i5 5 i 7 7 i9 2
21.31.
1 1 1 1 3 + + + + ⋅⋅⋅ = 1i 3 2 i 4 3 i5 4 i6 4
21.32.
1 1 1 1 π2 −8 2 + 2 2 + 2 2 + 2 2 + ⋅⋅⋅ = 16 1 i3 3 i5 5 i7 7 i9
21.33.
1 1 1 4π 2 − 39 2 2 + 2 2 2 + 2 2 2 + ⋅⋅⋅ = 16 1 i 2 i3 2 i3 i 4 3 i4 i5
21.34.
1 u a −1du 1 1 1 1 − + − + ⋅⋅⋅ = ∫ 0 1 + ud a a + d a + 2 d a + 3d
2
2
21.35.
2 2 p−1π 2 p Bp 1 1 1 1 2p + 2p + 2p + 2 p + ⋅⋅⋅ = (2 p)! 1 2 3 4
21.36.
(2 2 p − 1) π 2 p Bp 1 1 1 1 + + + + ⋅⋅⋅ = 2(2 p)! 12 p 32 p 5 2 p 7 2 p
21.37.
(2 2 p−1 − 1) π 2 p Bp 1 1 1 1 2p − 2p + 2p − 2 p + ⋅⋅⋅ = (2 p)! 1 2 3 4
21.38.
1 12 p+1
−
1 32 p+1
+
1 5 2 p+1
−
1 7 2 p+1
+ ⋅⋅⋅ =
π 2 p+1E p 2 2 p+ 2 (2 p)!
Miscellaneous Series 21.39.
1 sin(n + 1/2)α + cos α + cos 2α + ⋅⋅⋅ + cos nα = 2 2 sin(α /2)
21.40.
sin α + sin 2α + sin 3α + ⋅⋅⋅ + sin nα =
21.41.
1 + r cos α + r 2 cos 2α + r 3 cos 3α + ⋅⋅⋅ =
21.42.
r sin α + r 2 sin 2α + r 3 sin 3α + ⋅⋅⋅ =
21.43.
1 + r cos α + r 2 cos 2α + ⋅⋅⋅ + r n cos nα =
21.44.
r sin α + r 2 sin 2α + ⋅⋅⋅ + r n sin nα =
sin[1/2(n + 1)]α sin 1/2nα sin (α /2) 1 − r cos α , |r |< 1 1 − 2r cos α + r 2
r sin α , |r |< 1 1 − 2r cos α + r 2 r n+ 2 cos nα − r n+1 cos(n + 1)α − r cos α + 1 1 − 2 r cos α + r 2
r sin α − r n+1 sin(n + 1))α + r n+ 2 sin nα 1 − 2 r cos α + r 2
SERIES OF CONSTANTS
137
The Euler-Maclaurin Summation Formula n −1
21.45.
∑ F (k ) = ∫ k =1
n 0
1 F (k )dk − {F (0) + F (n)} 2
+
1 1 {F ′(n) − F (0)} − {F ′′′(n) − F ′′′(0)} 720 12
+
1 1 {F ( vii ) (n) − F ( vii ) (0)} {F ( v)) (n) − F ( v ) (0)} − 30, 240 1, 209, 600
+ ⋅⋅⋅ (−1) p−1
Bp {F ( 2 p−1) (n) − F ( 2 p−1) (0)} + ⋅⋅⋅ (2 p)!
The Poisson Summation Formula ∞
21.46.
∞
{
∑ F (k ) = ∑ ∫
k =−∞
m =−∞
∞
}
e 2π imx F ( x )dx
−∞
22
TAYLOR SERIES
Taylor Series for Functions of One Variable 22.1.
f ( x ) = f (a ) + f ′(a )( x − a ) +
f ′′(a )( x − a )2 f ( n−1) (a )( x − a ) + ⋅⋅⋅ + (n − 1)! 2!
n −1
+ Rn
where Rn, the remainder after n terms, is given by either of the following forms: 22.2. Lagrange’s form: Rn = 22.3. Cauchy’s form: Rn =
f ( n ) (ξ )( x − a )n n!
f ( n ) (ξ )( x − ξ )n−1 ( x − a ) (n − 1)!
The value x, which may be different in the two forms, lies between a and x. The result holds if f(x) has continuous derivatives of order n at least. If lim Rn = 0, the infinite series obtained is called the Taylor series for f(x) about x a. If a 0, the series n→∞ is often called a Maclaurin series. These series, often called power series, generally converge for all values of x in some interval called the interval of convergence and diverge for all x outside this interval. Some series contain the Bernoulli numbers Bn and the Euler numbers En defined in Chapter 23, pages 142143.
Binomial Series 22.4.
(a + x )n = a n + na n−1 x +
n(n − 1) n− 2 2 n(n − 1)(n − 2 ) n− 3 3 a x + a x + ⋅⋅⋅ 2! 3!
⎛ n⎞ ⎛ n⎞ ⎛ n⎞ = a n + ⎜ ⎟ a n−1 x + ⎜ ⎟ a n− 2 x 2 + ⎜ ⎟ a n− 3 x 3 + ⋅⋅⋅ ⎝ 3⎠ ⎝1⎠ ⎝ 2⎠
Special cases are 22.5.
(a + x )2 = a 2 + 2 ax + x 2
22.6.
(a + x )3 = a 3 + 3a 2 x + 3ax 2 + x 3
22.7.
(a + x )4 = a 4 + 4 a 3 x + 6 a 2 x 2 + 4 ax 3 + x 4
22.8.
(1 + x )−1 = 1 − x + x 2 − x 3 + x 4 − ⋅⋅⋅
1 < x < 1
22.9.
(1 + x )−2 = 1 − 2 x + 3x 2 − 4 x 3 + 5 x 4 − ⋅⋅⋅
1 < x < 1
(1 + x )−3 = 1 − 3x + 6 x 2 − 10 x 3 + 15 x 4 − ⋅⋅⋅
1 < x < 1
22.10.
138
TAYLOR SERIES
139
22.11.
(1 + x )−1/ 2 = 1 −
1 1i 3 2 1i 3i 5 3 x+ x − x + ⋅⋅⋅ 2 2i4 2i4i6
1 < x 1
22.12.
(1 + x )1/ 2 = 1 +
1 1 2 1i 3 3 x− x + x − ⋅⋅⋅ 2 2i4 2i4i6
1 < x 1
22.13.
(1 + x )−1/ 3 = 1 −
1 1i 4 2 1i 4 i 7 3 x+ x − x + ⋅⋅⋅ 3 3i6 3i6i9
1 < x 1
22.14.
(1 + x )1/ 3 = 1 +
1 2 2 2i5 3 x− x + x − ⋅⋅⋅ 3 3i6 3i6i9
1 < x 1
Series for Exponential and Logarithmic Functions x2 x3 + + ⋅⋅⋅ 2 ! 3!
∞ < x < ∞
22.15.
ex = 1 + x +
22.16.
a x = e x ln a = 1 + x ln a +
22.17.
ln (1 + x ) = x −
22.18.
1 ⎛1 + x⎞ x3 x5 x7 ln ⎜ =x+ + + + ⋅⋅⋅ ⎟ 2 ⎝1 − x⎠ 3 5 7
1 < x < 1
22.19.
3 5 1 ⎛ x − 1⎞ ⎪⎧⎛ x − 1⎞ 1 ⎛ x − 1⎞ ⎪⎫ ln x = 2 ⎨⎜ + ⋅⋅⋅⎬ + ⎜ + ⎜ ⎟ ⎟ ⎟ x + x x + 1 3 1 5 + 1 ⎝ ⎠ ⎝ ⎠ ⎝ ⎠ ⎪⎩ ⎪⎭
x>0
22.20.
⎛ x − 1⎞ 1 ⎛ x − 1⎞ 1 ⎛ x − 1⎞ + + + ⋅ ⋅⋅ ln x = ⎜ ⎝ x ⎟⎠ 2 ⎜⎝ x ⎟⎠ 3 ⎜⎝ x ⎟⎠
( x ln a )2 ( x ln a )3 + + ⋅⋅⋅ 2! 3!
x2 x3 x4 + − + ⋅⋅⋅ 2 3 4
2
∞ < x < ∞ 1 < x 1
3
x
1 2
Series for Trigonometric Functions 22.21.
sin x = x −
x3 x5 x7 + − + 3! 5! 7!
− ∞< x < ∞
22.22.
cos x = 1 −
x2 x4 x6 + − + 2! 4 ! 6 !
− ∞< x < ∞
22.23.
tan x = x +
2 2 n (2 2 n − 1)Bn x 2 n−1 x 3 2 x 5 17 x 7 + + + ++ 3 15 315 (2 n )!
| x |<
22.24.
cot x =
22.25.
sec x = 1 +
22.26.
csc x =
22.27.
sin −1 x = x +
22.28.
cos −1 x =
2 2 n Bn x 2 n−1 1 x x 3 2x5 − − − −− − x 3 45 945 (2 n )! E x 2n x 2 5 x 4 61x 6 + + ++ n + 2 24 720 (2 n )!
2(2 2 n−1 − 1)Bn x 2 n−1 1 x 7x 3 31x 5 + + + ++ + x 6 360 15, 120 (2 n )! 1 x3 1 i 3 x5 1 i 3 i 5 x7 + + + 2 3 2i4 5 2i4i6 7
1 x3 1 i 3 x5 π π ⎛ ⎞ − sin −1 x = − ⎜ x + + + ⎟ 2 2 ⎝ 2 3 2i4 5 ⎠
π 2
0< |x| <π |x|<
π 2
0< |x| <π |x|<1 |x|<1
TAYLOR SERIES
140
22.29.
22.30.
22.31. 22.32.
⎧ x3 x5 x7 x− + − + ⎪ 3 5 7 tan −1 x = ⎨ ⎪± π − 1 + 1 − 1 + ⎩ 2 x 3x 3 5x 5 ⎧π ⎛ ⎞ x3 x5 − ⎜ x − + − ⎟ ⎪ π ⎪ 3 5 ⎠ cot −1 x = − tan −1 x = ⎨ 2 ⎝ 2 ⎪ pπ + 1 − 1 + 1 − ⎪⎩ x 3x 3 5x 5
| x | <1 (+ if x 1, − if x − 1) | x | <1 ( p = 0 if x > 1, p = 1 if x < − 1)
⎞ π ⎛1 1 1⋅ 3 −⎜ + + + ⎟ 3 5 2 ⎝ x 2 i 3x 2 i 4 i 5x ⎠ 1 1 1 ⋅ 3 csc −1 x = sin −1 (1/ x ) = + + + x 2 i 3x 3 2 i 4 i 5x 5
sec −1 x = cos −1 (1/x ) =
|x|>1 |x|>1
Series for Hyperbolic Functions 22.33.
sinh x = x +
x3 x5 x7 + + + 3! 5! 7!
− ∞< x < ∞
22.34.
cosh x = 1 +
x2 x4 x6 + + + 2! 4 ! 6!
− ∞< x < ∞
22.35.
tanh x = x −
(−1)n−1 2 2 n (2 2 n − 1)Bn x 2 n−1 x 3 2 x 5 17 x 7 + − + + 3 15 315 (2 n )!
|x|<
22.36.
coth x =
22.37.
sech x = 1 −
22.38.
csch x =
22.39.
⎧ x3 1 i 3x 5 1 i 3 i 5x 7 ⎪⎪x − 2 i 3 + 2 i 4 i 5 − 2 i 4 i 6 i 7 + sinh −1 x = ⎨ ⎛ ⎞ ⎪± ln | 2 x | + 1 − 1 i 3 + 1 i 3 i 5 − 2 4 6 ⎜ ⎟⎠ 2 2 x 2 4 4 x 2 4 6 i i i i i i 6 x ⎩⎪ ⎝
22.40. 22.41. 22.42.
(−1)n−1 2 2 n Bn x 2 n−1 1 x x 3 2x5 + + − + + x 3 45 945 (2 n)! (−1)n En x 2 n x 2 5 x 4 61x 6 + − + + 2 24 720 (2 n )!
(−1)n 2(2 2 n−1 − 1)Bn x 2 n−1 1 x 7x 3 31x 5 + − + − + (2 n )! x 6 360 15, 120
⎛ 1 ⎞ ⎪⎫ 1i 3 1i 3i 5 ⎪⎧ + ⎟ ⎬ cosh −1 x = ± ⎨ln(2 x ) − ⎜ + + 2 4 6 2 i 4 i 4x 2 i 4 i 6 i 6x ⎝ 2 i 2x ⎠ ⎪⎭ ⎪⎩ 3 5 7 x x x tanh −1 x = x + + + + 3 5 7 coth −1 x =
1 1 1 1 + + + + x 3x 3 5 x 5 7 x 7
π 2
0< |x| <π |x|<
π 2
0< |x| <π | x | <1 ⎡ + if x 1 ⎤ ⎢− if x − 1⎥ ⎣ ⎦ ⎡+ if cosh −1 x > 0, x 1⎤ ⎢− if cosh −1 x < 0, x 1⎥ ⎣ ⎦
|x|<1 |x|>1
Miscellaneous Series x2 x4 x5 − − + 2 8 15
22.43.
e sin x = 1 + x +
22.44.
⎛ x 2 x 4 31x 6 ⎞ e cos x = e ⎜1 − + − + ⎟ 2 6 720 ⎝ ⎠
−∞< x < ∞ −∞< x < ∞
TAYLOR SERIES
141
x 2 x 3 3x 4 + + + 2 2 8
|x|<
π 2
22.45.
e tan x = 1 + x +
22.46.
e x sin x = x + x 2 +
2n / 2 sin (nπ /4) x n x3 x5 x6 + − − + + 3 30 90 n!
−∞< x < ∞
22.47.
e x cos x = 1 + x −
x3 x4 2 n /2 cos(nπ / 4 )x n − ++ + 3 6 n!
−∞< x < ∞
22.48.
ln | sin x | = ln | x | −
22.49.
ln | cos x | = −
22.50.
ln | tan x | = ln | x | +
22.51.
ln(1 + x ) = x − (1 + 12 ) x 2 + (1 + 12 + 13 ) x 3 − 1+ x
2 2 n−1 Bn x 2 n x2 x4 x6 + − − −− 6 180 2835 n(2 n )!
2 2 n−1 (2 2 n − 1)Bn x 2 n x 2 x 4 x 6 17 x 8 + − − − −− 2 12 45 2520 n(2 n )! 2 2 n (2 2 n−1 − 1)Bn x 2 n x 2 7 x 4 62 x 6 + + + ++ 3 90 2835 n(2 n )!
0< |x| <π |x|<
π 2
0< |x|<
π 2
| x | <1
Reversion of Power Series Suppose 22.52.
y = C1 x + C2 x 2 + C3 x 3 + C4 x 4 + C5 x 5 + C6 x 6 +
then 22.53.
x = C1 y + C2 y 2 + C3 y 3 + C4 y 4 + C5 y 5 + C6 y 6 +
where 22.54.
c1C1 = 1
22.55.
c13C2 = −c2
22.56.
c15C3 = 2c22 − c1c3
22.57.
c17C4 = 5c1c2 c3 − 5c23 − c12 c4
22.58.
c19C5 = 6c12 c2 c4 + 3c12 c32 − c13c5 + 14 c24 − 21c1c22 c3
22.59.
c111C6 = 7c13c2 c5 + 84 c1c23c3 + 7c13c3c4 − 28c12 c2 c32 − c14 c6 − 28c12 c22 c4 − 42c25
Taylor Series for Functions of Two Variables 22.60.
f ( x, y) = f (a, b ) + ( x − a ) f x (a, b ) + ( y − b ) f y (a, b ) +
1 {( x − a )2 f xx (a, b ) + 2( x − a )( y − b ) f xy (a, b ) + ( y − b )2 f yy (a, b )} + 2!
where f x (a, b ), f y (a, b ),… denote partial derivatives with respect to x, y, … evaluated at x a, y b.
23
BERNOULLI and EULER NUMBERS
Definition of Bernoulli Numbers The Bernoulli numbers B1 , B2 , B3 ,… are defined by the series
23.1.
x x B x2 B x4 B x6 = 1− + 1 − 2 + 3 − 2 2! 4! 6! e −1
23.2.
1−
x
x x B x2 B x4 B x6 cot = 1 + 3 + 3 + 2 2 2! 4! 6!
| x | < 2π |x| <π
Definition of Euler Numbers The Euler numbers E1, E2, E3, … are defined by the series
23.3.
sech x = 1 −
23.4.
sec x = 1 +
E1 x 2 E2 x 4 E3 x 6 + − + 2! 4! 6!
E1 x 2 E2 x 4 E3 x 6 + − + 2! 4! 6!
|x|<
π 2
|x|<
π 2
Table of First Few Bernoulli and Euler Numbers Bernoulli Numbers
142
Euler Numbers
B1 = 1/6
E1 = 1
B2 = 1/30
E2 = 5
B3 = 1/42
E3 = 61
B4 = 1/30
E4 = 1385
B5 = 5/66
E5 = 50, 521
B6 = 691/2730
E6 = 2, 702, 765
B7 = 7/6
E7 = 199, 360, 981
B8 = 3617/510
E8 = 19, 391, 512, 145
B9 = 43, 867/798
E9 = 2, 404, 879, 675, 441
B10 = 174, 611/330
E10 = 370, 371, 188, 237, 525
B11 = 854, 513138 /
E11 = 69, 348, 874, 393, 137, 901
B12 = 236, 364, 091/2730
E12 = 15, 514, 534, 163, 557, 086, 905
BERNOULLI AND EULER NUMBERS
Relationships of Bernoulli and Euler Numbers 23.5.
⎛ 2 n + 1⎞ 2 ⎛ 2 n + 1⎞ 4 ⎛ 2 n + 1⎞ 6 n −1 2n ⎜⎝ 2 ⎟⎠ 2 B1 − ⎜⎝ 4 ⎟⎠ 2 B2 + ⎜⎝ 6 ⎟⎠ 2 B3 − (−1) (2 n + 1)2 Bn = 2 n
23.6.
⎛ 2 n⎞ ⎛ 2 n⎞ ⎛ 2 n⎞ En = ⎜ ⎟ En−1 − ⎜ ⎟ En− 2 + ⎜ ⎟ En− 3 − (−1)n ⎝ 2⎠ ⎝ 4⎠ ⎝ 6⎠
23.7.
Bn =
⎫ ⎧⎛ 2 n − 1⎞ 2n ⎛ 2 n − 1⎞ ⎛ 2 n − 1⎞ E − E + E − (−1)n−1 ⎬ 2 2 n (2 2 n − 1) ⎨⎩⎜⎝ 1 ⎟⎠ n−1 ⎜⎝ 3 ⎟⎠ n− 2 ⎜⎝ 5 ⎟⎠ n− 3 ⎭
Series Involving Bernoulli and Euler Numbers 23.8. 23.9. 23.10. 23.11.
1 1 ⎫ ⎧ ⎨1 + 2 2 n + 32 n + ⎬ ⎭ ⎩ 2(2 n )! ⎧ 1 1 ⎫ 1+ Bn = 2 n + + ⎬ (2 − 1)π 2 n ⎨⎩ 32 n 5 2 n ⎭ Bn =
(2 n )! 2 2 n−1 π 2 n
2(2 n )! 1 1 ⎫ ⎧ + − ⎬ 1− (2 2 n−1 − 1)π 2 n ⎨⎩ 2 2 n 32 n ⎭ 2 n+ 2 2 (2 n )! ⎧ 1 1 ⎫ 1− En = + − ⎬ π 2 n+1 ⎨⎩ 32 n+1 5 2 n+1 ⎭ Bn =
Asymptotic Formula for Bernoulli Numbers 23.12.
Bn ~ 4 n 2 n (π e)−2 n π n
143
24
FOURIER SERIES
Definition of a Fourier Series The Fourier series corresponding to a function f (x) defined in the interval c x c + 2 L where c and L > 0 are constants, is defined as
24.1.
∞ a0 nπx n π x⎞ ⎛ + ∑ ⎜ an cos + bn sin 2 n=1 ⎝ L L ⎟⎠
where
24.2.
⎧ ⎪an = ⎨ ⎪bn = ⎩
1 L 1 L
∫ ∫
c+ 2 L
c c+ 2 L c
nπx dx L nπx f ( x )sin dx L
f ( x ) cos
If f (x) and f ′(x) are piecewise continuous and f (x) is defined by periodic extension of period 2L, i.e., f (x 2L) f (x), then the series converges to f (x) if x is a point of continuity and to 12 { f ( x + 0) + f ( x − 0)} if x is a point of discontinuity.
Complex Form of Fourier Series Assuming that the series 24.1 converges to f (x), we have
24.3.
f (x ) =
∞
∑c e
in π x/L
n
n =−∞
where 24.4.
1 cn = 2L
∫
c+ 2 L c
f ( x )e
− in π x/L
⎧12 (an − ibn ) ⎪ dx = ⎨12 (a− n + ib− n ) ⎪⎩12 a0
n>0 n<0 n=0
Parseval’s Identity 24.5.
1 L
∫
c+ 2 L c
{ f ( x )}2 dx =
∞ a02 + ∑ (an2 + bn2 ) 2 n=1
Generalized Parseval Identity 24.6.
1 L
∫
c+ 2 L c
f ( x )g( x ) dx =
∞ a0 c0 + ∑ (an cn + bn d n ) 2 n =1
where an, bn and cn, dn are the Fourier coefficients corresponding to f (x) and g(x), respectively.
144
FOURIER SERIES
145
Special Fourier Series and Their Graphs
24.7.
⎧ 1 0< x<π f (x ) = ⎨ ⎩−1 −π < x < 0 4 ⎛ sin x sin 3x sin 5 x ⎞ + + + ⎟ 3 5 π ⎜⎝ 1 ⎠ Fig. 24-1
24.8.
0< x<π ⎧ x f (x ) = | x | = ⎨ π − x − <x<0 ⎩
π 4 ⎛ cos x cos 3x cos 5 x ⎞ − + + + ⎟ 2 π ⎜⎝ 12 32 52 ⎠
Fig. 24-2
24.9.
f ( x ) = x, − π < x < π ⎛ sin x sin 2 x sin 3x ⎞ 2⎜ − + − ⎟ 2 3 ⎝ 1 ⎠
Fig. 24-3
24.10.
f ( x ) = x, 0 < x < 2π ⎛ sin x sin 2 x sin 3x ⎞ π − 2⎜ + + + ⎟ 2 3 ⎝ 1 ⎠
Fig. 24-4
24.11.
f ( x ) = | sin x |, − π < x < π 2 4 ⎛ cos 2 x cos 4 x cos 6 x ⎞ − + + + ⎟ 3i5 5i7 π π ⎜⎝ 1 i 3 ⎠
Fig. 24-5
FOURIER SERIES
146
24.12.
⎧sin x 0 < x < π f (x ) = ⎨ π < x < 2π ⎩ 0 1 1 2 ⎛ cos 2 x cos 4 x cos 6 x ⎞ + sin x − ⎜ + + + ⎟ 3i5 5i7 π 2 π ⎝ 1i 3 ⎠
Fig. 24-6
24.13.
0< x<π ⎧ cos x f (x ) = ⎨ π − cos x − <x<0 ⎩ 8 ⎛ sin 2 x 2 sin 4 x 3 sin 6 x ⎞ + + + ⎟ 3i5 5i7 π ⎜⎝ 1 i 3 ⎠
Fig. 24-7
24.14.
f (x ) = x 2 , − π < x < π
π2 ⎛ cos x cos 2 x cos 3x ⎞ − 4⎜ 2 − + − ⎟ 3 22 32 ⎝ 1 ⎠
Fig. 24-8
24.15.
f ( x ) = x (π − x ), 0 < x < π
π 2 ⎛ cos 2 x cos 4 x cos 6 x ⎞ − + + + ⎟ 6 ⎜⎝ 12 22 32 ⎠
Fig. 24-9
24.16.
f ( x ) = x (π − x )(π + x ), − π < x < π ⎛ sin x sin 2 x sin 3x ⎞ + − ⎟ 12 ⎜ 3 − 23 33 ⎝ 1 ⎠
Fig. 24-10
FOURIER SERIES
24.17.
147
0 < x < π −α ⎧0 ⎪ f ( x ) = ⎨1 π − α < x < π + α ⎪⎩0 π + α < x < 2π
α 2 ⎛ sin α cos x sin 2α cos 2 x − − π π ⎜⎝ 1 2 +
24.18.
sin 3α cos 3x ⎞ − ⎟ 3 ⎠
Fig. 24-11
0< x<π ⎧ x (π − x ) f (x ) = ⎨ π π − x ( − x ) − <x<0 ⎩ 8 ⎛ sin x sin 3x sin 5 x ⎞ + + + ⎟ π ⎜⎝ 13 33 53 ⎠
Fig. 24-12
Miscellaneous Fourier Series 24.19.
f ( x ) = sin μ x, − π < x < π , μ ≠ integer 2 sin μπ ⎛ sin x 2 sin 2 x 3 sin 3x ⎞ + − ⎟ ⎜ 2 2 − 2 π 2 − μ 2 32 − μ 2 ⎝1 − μ ⎠
24.20.
f ( x ) = cos μ x, − π < x < π , μ ≠ integer cos 2 x cos 3x 2 μ sin μπ ⎛ 1 cos x ⎞ ⎜⎝ 2 μ 2 + 12 − μ 2 − 2 2 − μ 2 + 32 − μ 2 − ⎟⎠ π
24.21.
f ( x ) = tan −1[(a sin x ) / (1 − a cos x )], − π < x < π , | a | < 1 a sin x +
24.22.
a2 a3 sin 2 x + sin 3x + 2 3
f ( x ) = ln (1 − 2a cos x + a 2 ), − π < x < π , | a | < 1 ⎛ ⎞ a3 a2 − 2 ⎜ a cos x + cos 2 x + cos 3x + ⎟ 2 3 ⎝ ⎠
24.23.
f (x) =
1 −1 tan [(2a sin x ) / (1 − a 2 )], − π < x < π , | a | < 1 2 a sin x +
24.24.
f (x) =
a3 a5 sin 3x + sin 5x + 3 5
1 −1 tan [(2a cos x ) / (1 − a 2 )], − π < x < π , | a | < 1 2 a cos x −
a3 a5 cos 3x + cos 5x − 3 5
FOURIER SERIES
148
24.25.
f (x ) = e μx , − π < x < π ∞ 2 sinh μπ ⎛ 1 (−1)n ( μ cos nx − n sin nx )⎞ +∑ ⎜ ⎟⎠ π μ 2 + n2 ⎝ 2 μ n=1
24.26.
f ( x ) = sinh μ x, − π < x < π 2 sinh μπ ⎛ sin x 2 sin 2 x 3 sin 3x ⎞ + − ⎟ ⎜ 2 2 − 2 π 2 + μ 2 32 + μ 2 ⎝1 + μ ⎠
24.27.
f ( x ) = cosh μ x, − π < x < π 2 μ sinh μπ ⎛ 1 cos 3x cos x cos 2 x ⎞ ⎜⎝ 2 μ 2 − 12 + μ 2 + 2 2 + μ 2 − 32 + μ 2 + ⎟⎠ π
24.28.
f ( x ) = ln | sin 12 x |, 0 < x < π cos x cos 2 x cos 3x ⎛ ⎞ + ⎟ − ⎜ ln 2 + + + 3 1 2 ⎝ ⎠
24.29.
f ( x ) = ln | cos 12 x |, − π < x < π cos x cos 2 x cos 3x ⎛ ⎞ + ⎟ − ⎜ ln 2 − + − 3 1 2 ⎝ ⎠
24.30.
f ( x ) = 16 π 2 − 12 π x + 14 x 2 , 0 x 2π cos x cos 2 x cos 3x + + + 32 12 22
24.31.
f (x ) =
1 12
x ( x − π )( x − 2π ), 0 x 2π sin x sin 2 x sin 3x + + + 33 13 23
24.32.
f (x ) =
1 90
π 4 − 121 π 2 x 2 + 121 π x 3 −
1 48
x 4 , 0 x 2π cos x cos 2 x cos 3x + + + 24 34 14
Section VII: Special Functions and Polynomials
25
THE GAMMA FUNCTION
Definition of the Gamma Function ⌫(n) for n > 0 25.1.
∞
Γ(n) = ∫ t n−1e − t dt 0
n>0
Recursion Formula 25.2.
Γ (n + 1) = nΓ (n)
If n = 0, 1, 2, …, a nonnegative integer, we have the following (where 0! = 1): 25.3.
Γ(n + 1) = n!
The Gamma Function for n < 0 For n < 0 the gamma function can be defined by using 25.2, that is, 25.4.
Γ (n) =
Γ (n + 1) n
Graph of the Gamma Function
5
Γ(n)
4 3 2 1 −5
−4
−3
−2
−1
1 −1
2
3
4
5
n
−2 −3 −4 −5
Fig. 25-1
149 Copyright © 2009, 1999, 1968 by The McGraw-Hill Companies, Inc. Click here for terms of use.
THE GAMMA FUNCTION
150
Special Values for the Gamma Function 25.5.
Γ( 12 ) = π
25.6.
Γ(m + 12 ) =
25.7.
Γ(− m + 12 ) =
1i 3 i 5 ⋅⋅⋅ (2m − 1) 2m
π
(−1)m 2m π 1i 3 i 5 ⋅⋅⋅ (2m − 1)
m = 1, 2, 3, ... m = 1, 2, 3, ...
Relationships Among Gamma Functions π sin pπ
25.8.
Γ ( p)Γ (1− p) =
25.9.
22 x −1 Γ ( x )Γ ( x + 12 ) = π Γ (2 x )
This is called the duplication formula. 25.10.
⎛ ⎛ m − 1⎞ 1⎞ ⎛ 2⎞ Γ ( x )Γ ⎜ x + ⎟ Γ ⎜ x + ⎟ ⋅⋅⋅ Γ ⎜ x + = m1/ 2 − mx (2π )( m −1)/ 2 Γ (mx ) m m m ⎟⎠ ⎝ ⎠ ⎝ ⎠ ⎝
For m = 2 this reduces to 25.9.
Other Definitions of the Gamma Function 1 i 2 i 3 ⋅⋅⋅ k kx ( x + 1)( x + 2) ⋅⋅⋅ ( x + k )
25.11.
Γ( x + 1) = lim
25.12.
∞ ⎧ ⎫⎪ 1 x⎞ ⎪⎛ = xeγ x ∏ ⎨⎜1 + ⎟ e − x / m ⎬ Γ( x ) ⎪⎭ m =1 ⎪ ⎩⎝ m ⎠
k →∞
This is an infinite product representation for the gamma function where g is Euler’s constant defined in 1.3, page 3.
Derivatives of the Gamma Function ∞
25.13.
Γ ′(1) = ∫ e − x ln x dx = −γ
25.14.
⎛1 Γ ′( x ) = −γ + ⎜ − Γ(x) ⎝1
0
⎛1 1⎞ ⎛ 1 1 ⎞ 1 ⎞ +⎜ − + ⋅⋅⋅ + ⋅⋅⋅ + ⎜ − ⎟ ⎟ x ⎠ ⎝ 2 x + 1⎠ ⎝ n x + n − 1⎟⎠
Here again is Euler’s constant g.
THE GAMMA FUNCTION
151
Asymptotic Expansions for the Gamma Function 25.15.
⎧ ⎫ 1 1 139 Γ( x + 1) = 2π x x x e − x ⎨1 + + − + ⋅⋅ ⋅⎬ 2 3 x x , x 12 288 51 840 ⎩ ⎭
This is called Stirling’s asymptotic series. If we let x = n a positive integer in 25.15, then a useful approximation for n! where n is large (e.g., n > 10) is given by Stirling’s formula 25.16.
n! ~ 2π n n ne − n
where ~ is used to indicate that the ratio of the terms on each side approaches 1 as n → ∞.
Miscellaneous Results 25.17.
| Γ (ix ) |2 =
π x sinh π x
26
THE BETA FUNCTION
Definition of the Beta Function B(m, n) 26.1.
1
B(m, n) = ∫ t m −1 (1 − t )n−1 dt 0
m > 0, n > 0
Relationship of Beta Function to Gamma Function 26.2.
B( m , n ) =
Γ (m )Γ (n) Γ (m + n)
Extensions of B(m, n) to m < 0, n < 0 are provided by using 25.4.
Some Important Results 26.3.
B( m , n ) = B( n , m )
26.4.
B( m , n ) = 2 ∫
26.5.
B( m , n ) = ∫
26.6.
B(m, n) = r n (r + 1)m
152
π /2
0
∞
0
sin 2 m −1 θ cos 2 n−1 θ dθ
t m −1 dt (1 + t )m + n
∫
(1 − t )n−1 dt (r + t )m + n
1 m −1
0
t
27
BESSEL FUNCTIONS
Bessel’s Differential Equation 27.1.
x 2 y n + xy′ + ( x 2 − n 2 ) y = 0
n0
Solutions of this equation are called Bessel functions of order n.
Bessel Functions of the First Kind of Order n 27.2.
⎧ ⎫ xn x2 x4 + − ⋅⋅⋅⎬ ⎨1 − 2 Γ (n + 1) ⎩ 2(2n + 2) 2 i 4(2n + 2)(2n + 4) ⎭
Jn ( x ) =
n
∞
(−1) k ( x /2)n+ 2 k k ! Γ (n + k + 1)
=∑ k =0
27.3.
J− n ( x ) =
x −n x2 x4 ⎫ ⎧ − ⋅⋅⋅⎬ 1− + ⎨ 2 Γ (1 − n) ⎩ 2(2 − 2n) 2 i 4(2 − 2n)(4 − 2n) ⎭ −n
∞
=∑ k =0
27.4.
(−1) k ( x / 2)2 k − n k ! Γ ( k + 1 − n)
J − n ( x ) = (−1)n J n ( x )
n = 0, 1, 2, ...
If n ≠ 0, 1, 2, ..., J n ( x ) and J–n (x) are linearly independent. If n ≠ 0, 1, 2, ..., J n ( x ) is bounded at x = 0 while J–n (x) is unbounded. For n = 0, 1 we have
x2 x4 x6 + − + ⋅⋅⋅ 22 22 i 4 2 22 i 4 2 i62
27.5.
J0 ( x ) = 1 −
27.6.
J1 ( x ) =
27.7.
J 0(x) = − J1 (x)
x x3 x5 x7 − 2 + 2 2 2 − 2 2 2 + ⋅⋅⋅ 2 2 i4 2 i4 i6 2 i 4 i6 i8
Bessel Functions of the Second Kind of Order n
27.8.
⎧ J n ( x ) cos nπ − J − n ( x ) ⎪ sin n π ⎪ Yn ( x ) = ⎨ J p ( x ) cos pπ − J − p ( x ) ⎪ ⎪lim p→ n sin pπ ⎩
n ≠ 0, 1, 2, ... n = 0, 1, 2, ...
This is also called Weber’s function or Neumann’s function [also denoted by Nn(x)].
153
BESSEL FUNCTIONS
154 For n = 0, 1, 2, …, L’ Hospital’s rule yields 27.9.
Yn ( x ) =
2 1 n −1 (n − k − 1)! {ln ( x / 2) + γ }J n ( x ) − ∑ ( x / 2)2 k − n π π k =0 k! −
1 ∞ ( x / 2)2 k + n (−1) k {Φ(k ) + Φ(n + k )} ∑ π k =0 k !(n + k )!
where g = .5772156 … is Euler’s constant (see 1.20) and 27.10.
Φ( p) = 1 +
1 1 1 + + ⋅⋅⋅ + , P 2 3
Φ (0) = 0
For n = 0, 2 2 {ln ( x / 2) + γ }J 0 ( x ) + π π
27.11.
Y0 ( x ) =
27.12.
Y− n ( x ) = (−1)n Yn ( x )
( )
(
n = 0, 1, 2, ...
For any value n 0, J n ( x ) is bounded at x = 0 while Yn(x) is unbounded.
General Solution of Bessel’s Differential Equation 27.13.
y = AJ n ( x ) + BJ − n ( x )
n ≠ 0, 1, 2, .. .
27.14.
y = AJ n ( x ) + BYn ( x )
all n
27.15.
y = AJ n ( x ) + BJ n ( x ) ∫
dx xJ n2 ( x )
all n
where A and B are arbitrary constants.
Generating Function for Jn(x) 27.16.
e x ( t −1/t ) / 2 =
∞
∑ J ( x )t
n
n
n =−∞
Recurrence Formulas for Bessel Functions 2n J ( x ) − J n−1 ( x ) x n
27.17.
J n+1 ( x ) =
27.18.
J n′ ( x ) = 12 {J n −1 ( x ) − J n+1 ( x )}
27.19.
xJ n′ ( x ) = xJ n−1 ( x ) − nJ n ( x )
27.20.
xJ n′ ( x ) = nJ n ( x ) − xJ n+1 ( x )
)
1 1 1 x4 x6 ⎫ ⎧x 2 ⎨ 22 − 22 4 2 1 + 2 + 22 4 2 62 1 + 2 + 3 − ⋅⋅⋅⎬ ⎭ ⎩
BESSEL FUNCTIONS
27.21.
d n {x J n ( x )} = x n J n−1 ( x ) dx
27.22.
d −n {x J n ( x )} = − x − n J n+1 ( x ) dx
155
The functions Yn(x) satisfy identical relations.
Bessel Functions of Order Equal to Half an Odd Integer In this case the functions are expressible in terms of sines and cosines. 2 sin x πx
27.23.
J1/ 2 ( x ) =
27.24.
J −1/ 2 ( x ) =
27.25.
J3 / 2 ( x ) =
2 cos x πx 2 πx
⎛ sin x ⎞ ⎜⎝ x − cos x⎟⎠
⎞ 2 ⎛ cos x + sin x⎟ π x ⎜⎝ x ⎠
27.26.
J −3 / 2 ( x ) =
27.27.
J5 / 2 ( x ) =
⎫⎪ ⎞ 2 ⎧⎪ ⎛ 3 3 ⎨ ⎜ 2 − 1⎟ sin x − cos x⎬ π x ⎩⎪ ⎝ x x ⎠ ⎪⎭
27.28.
J −5 / 2 ( x ) =
⎛3 ⎞ 2 ⎧⎪ 3 ⎪⎫ ⎨ sin x + ⎜ − 1⎟ cos x⎬ π x ⎩⎪ x ⎝x ⎠ ⎭⎪
For further results use the recurrence formula. Results for Y1/ 2 ( x ), Y3 / 2 ( x ), ... are obtained from 27.8.
Hankel Functions of First and Second Kinds of Order n 27.29.
H n(1) ( x ) = J n ( x ) + iYn ( x )
27.30.
H n( 2) ( x ) = J n ( x ) − iYn ( x )
Bessel’s Modified Differential Equation 27.31.
x 2 y′′ + xy′ − ( x 2 + n 2 ) y = 0
n0
Solutions of this equation are called modified Bessel functions of order n.
Modified Bessel Functions of the First Kind of Order n 27.32.
I n ( x ) = i − n J n (ix ) = e − nπ i / 2 J n (ix ) =
27.33.
x4 ( x / 2)n+ 2 k xn x2 ⎫ ∞ ⎧ 1 + + ⋅⋅⋅ + =∑ ⎬ ⎨ n 2 Γ (n + 1) ⎩ 2(2n + 2) 2 i 4(2n + 2)(2n + 4) ⎭ k = 0 k ! Γ (n + k + 1)
I − n ( x ) = i n J − n (ix ) = e nπ i / 2 J − n (ix ) =
27.34.
⎧ ⎫ ∞ x −n x2 x4 ( x / 2)2 k − n 1 + + + ⋅⋅⋅ = ⎨ ⎬ ∑ 2 − n Γ (1 − n) ⎩ 2(2 − 2n) 2 i 4(2 − 2n)(4 − 2n) ⎭ k = 0 k ! Γ ( k + 1 − n)
I − n ( x ) = I n ( x ) n = 0, 1, 2, ...
156
BESSEL FUNCTIONS
If n ≠ 0, 1, 2, ..., then In(x) and I–n(x) are linearly independent. For n = 0, 1, we have x2 x4 x6 + + + ⋅⋅⋅ 22 22 i 4 2 22 i 4 2 i 62
27.35.
I 0 (x) = 1 +
27.36.
I1 ( x ) =
27.37.
I 0′ ( x ) = I1 ( x )
x x3 x5 x7 + 2 + 2 2 + 2 2 2 + ⋅⋅⋅ 2 2 i 4 2 i 4 i6 2 i 4 i 6 i8
Modified Bessel Functions of the Second Kind of Order n
27.38.
⎧ π ⎪ 2 sin nπ {I − n ( x ) − I n ( x )} ⎪ K n (x) = ⎨ ⎪lim π {I ( x ) − I p ( x )} ⎪⎩p→n 2 sin pπ − p
n ≠ 0, 1, 2, ... n = 0, 1, 2, ...
For n = 0, 1, 2, …, L’ Hospital’s rule yields 27.39.
K n ( x ) = (−1)n+1{ln ( x / 2) + γ )I n ( x ) + +
(−1)n 2
∞
1 n−1 (−1) k (n − k − 1)!( x / 2)2 k − n 2∑ k =0
( x / 2)n+ 2 k
∑ k !(n + k )! {Φ(k ) + Φ(n + k )} k =0
where Φ(p) is given by 27.10. For n = 0, 27.40.
K 0 ( x ) = −{ln ( x / 2) + γ }I 0 ( x ) +
27.41.
K −n (x) = K n (x)
⎛ 1 1⎞ x2 x 4 ⎛ 1⎞ x6 + 2 2 ⎜1 + ⎟ + 2 2 2 ⎜1 + + ⎟ + ⋅⋅⋅ 2 2 2 i 4 ⎝ 2⎠ 2 i 4 i 6 ⎝ 2 3⎠
n = 0, 1, 2, ...
General Solution of Bessel’s Modified Equation 27.42.
y = AI n ( x ) + BI − n ( x )
n ≠ 0, 1, 2, ...
27.43.
y = AI n ( x ) + BK n ( x )
all n
27.44.
y = AI n ( x ) + BI n ( x )
dx 2 n (x)
∫ xI
where A and B are arbitrary constants.
Generating Function for In(x) 27.45.
e x ( t +1/ t ) / 2 =
∞
∑ I ( x )t n
n =−∞
n
all n
BESSEL FUNCTIONS
157
Recurrence Formulas for Modified Bessel Functions 27.46.
I n+1 ( x ) = I n−1 ( x ) −
2n I (x) x n
27.52.
K n+1 ( x ) = K n−1 ( x ) +
2n K (x) x n
27.47.
I n′ ( x ) = 12 {I n−1 ( x ) + I n+1 ( x )}
27.53.
K n′ ( x ) = − 12 {K n−1 ( x ) + K n+1 ( x )}
27.48.
xI n′ ( x ) = xI n −1 ( x ) − nI n ( x )
27.54.
xK n′ ( x ) = − xK n−1 ( x ) − nK n ( x )
27.49.
xI n′ ( x ) = xI n+1 ( x ) + nI n ( x )
27.55.
xK n′ ( x ) = nK n ( x ) − xK n+1 ( x )
27.50.
d n {x I n ( x )} = x n I n−1 ( x ) dx
27.56.
d n {x K n ( x )} = − x n K n−1 ( x ) dx
27.51.
d −n {x I n ( x )} = x − n I n+1 ( x ) dx
27.57.
d −n {x K n ( x )} = − x − n K n+1 ( x ) dx
Modified Bessel Functions of Order Equal to Half an Odd Integer In this case the functions are expressible in terms of hyperbolic sines and cosines. 27.58.
I1 / 2 ( x ) =
27.59.
I −1/ 2 ( x ) =
27.60.
I3/ 2 (x) =
2 sinh x πx 2 cosh x πx 2 ⎛ sinh x ⎞ cosh x − x ⎟⎠ π x ⎜⎝
27.61.
I −3 / 2 ( x ) =
27.62.
I5/ 2 (x) =
27.63.
I −5 / 2 ( x ) =
2 πx 2 πx 2 πx
cosh x ⎞ ⎛ ⎜⎝sinh x − x ⎟⎠ 3 ⎫ ⎧⎛ 3 ⎞ ⎨⎜⎝ x 2 + 1⎟⎠ sinh x − x cosh x⎬ ⎭ ⎩ 3 ⎫ ⎧⎛ 3 ⎞ ⎨⎜⎝ x 2 + 1⎟⎠ cosh x − x sinh x⎬ ⎭ ⎩
For further results use the recurrence formula 27.46. Results for K1/2(x), K3/2(x), … are obtained from 27.38.
Ber and Bei Functions The real and imaginary parts of J n ( xe3π i / 4 ) are denoted by Bern (x) and Bein(x) where ( x / 2)2 k + n (3n + 2k )π cos 1 k ! ( n + k + ) Γ 4 k =0 ∞
27.64.
Bern ( x ) = ∑
27.65.
Bei n ( x ) = ∑
( x / 2)2 k + n (3n + 2k )π sin 1 k ! ( n + k + ) Γ 4 k =0 ∞
If n = 0. ( x / 2)4 ( x / 2)8 + − ⋅⋅⋅ 4!2 2!2
27.66.
Ber ( x ) = 1 −
27.67.
Bei( x ) = ( x / 2)2 −
( x / 2)6 ( x / 2)10 + − ⋅⋅⋅ 3!2 5!2
BESSEL FUNCTIONS
158
Ker and Kei Functions The real and imaginary parts of e − nπ i / 2 K n ( xeπ i / 4 ) are denoted by Kern(x) and Kein(x) where 27.68.
27.69.
Kern ( x ) = −{ln ( x / 2) + γ }Bern ( x ) + 14 π Bei n ( x ) +
1 n −1 (n − k − 1)!( x / 2)2 k − n (3n + 2k )π cos 2∑ k ! 4 k =0
+
1 2
∞
( x / 2)n+ 2 k
∑ k !(n + k )! {Φ(k ) + Φ(n + k )}cos k =0
(3n + 2k)π 4
Kei n ( x ) = −{ln ( x / 2) + γ }Bei n ( x ) − 14 π Bern ( x ) −
1 n −1 (n − k − 1)!( x / 2)2 k − n (3n + 2 k )π sin ∑ 2 k =0 k! 4
+
1 2
∞
∑ k =0
( x / 2)n+ 2 k (3n + 2k)π {Φ(k ) + Φ(n + k )}sin k !(n + k )! 4
and Φ is given by 27.10. If n = 0, 27.70.
Ker ( x ) = −{ln ( x / 2) + γ }Ber ( x ) +
π ( x / 2)4 ( x / 2)8 1 Bei( x ) + 1 − ( 1 + ) + (1 + 12 + 13 + 14 ) − ⋅⋅⋅ 2 4 2!2 4!2
27.71.
Kei( x ) = −{ln ( x / 2) + γ }Bei( x ) −
π ( x / 2)6 (1 + 12 + 13 ) + ⋅⋅⋅ Ber(x ) + ( x / 2)2 − 3!2 4
Differential Equation For Ber, Bei, Ker, Kei Functions 27.72.
x 2 y′′ + xy′ − (ix 2 + n 2 ) y = 0
The general solution of this equation is 27.73.
y = A{Bern ( x ) + i Bei n ( x )} + B{Kern ( x ) + i Kei n ( x )}
Graphs of Bessel Functions
Fig. 27-1
Fig. 27-2
BESSEL FUNCTIONS
159
Fig. 27-3
Fig. 27-4
Fig. 27-5
Fig. 27-6
Indefinite Integrals Involving Bessel Functions 27.74.
∫ xJ ( x )dx = xJ ( x )
27.75.
∫ x J ( x )dx = x J ( x) + xJ ( x) − ∫ J ( x )dx
27.76.
∫x
27.77.
∫
J0 ( x ) J (x) dx = J1 ( x ) − 0 − ∫ J 0 ( x )dx 2 x x
27.78.
∫
J0 ( x ) J0 ( x ) J1 ( x ) 1 dx = − − xm (m − 1)2 x m − 2 (m − 1) x m −1 (m − 1)2
27.79.
∫ J ( x )dx = − J ( x)
27.80.
∫ xJ ( x)dx = − xJ ( x) + ∫ J ( x)dx
27.81.
∫x
0
1
2
2
0
m
1
0
0
J 0 ( x )dx = x m J1 ( x ) + (m − 1) x m −1J 0 ( x ) − (m − 1)2
1
0
1
m
0
0
J1 ( x )dx = − x m J 0 ( x ) + m
∫x
m −1
J 0 ( x )dx
∫x
∫
m−2
J 0 ( x )dx
J0 ( x ) dx x m−2
BESSEL FUNCTIONS
160
27.82.
∫
J1 ( x ) dx = − J1 ( x ) + ∫ J 0 ( x )dx x
27.83.
∫
J1 ( x ) J1 ( x ) 1 + m dx = − x mx m −1 m
27.84.
∫x
27.85.
∫x
27.86.
∫x
27.87.
∫ xJ (α x ) J (β x)dx =
27.88.
∫ xJ
n
∫
J0 ( x ) dx x m −1
J n−1 ( x )dx = x n J n ( x )
−n
m
J n+1 ( x )dx = − x − n J n ( x )
J n ( x )dx = − x m J n−1 ( x ) + (m + n − 1)
n
2 n
n
(α x )dx =
∫x
m −1
J n−1 ( x )dx
x{α J n (β x ) J n′ (α x ) − β J n (α x ) J n′ (β x )} β2 −α2
x2 x2 ⎛ n2 ⎞ {J n′ (α x )}2 + ⎜1 − 2 2 ⎟ {J n (α x )}2 2 2 ⎝ α x ⎠
The above results also hold if we replace Jn(x) by Yn(x) or, more generally, AJn(x) + BYn(x) where A and B are constants.
Definite Integrals Involving Bessel Functions 27.89.
∫
27.90.
∫
∞ 0
∞ 0
∞
∫
0
27.92.
∫
0
27.93.
∫
0
27.94.
∫
27.95.
∫
0
27.96.
∫
0
27.97.
∫
27.91.
∞
∞
∞
0 1
1
1
0
1
e − ax J 0 (bx )dx = e − ax J n (bx )dx =
a + b2 2
( a 2 + b 2 − a) n bn a2 + b2
⎧ 1 ⎪ 2 cos ax J 0 (bx )dx = ⎨ a − b 2 ⎪ 0 ⎩ J n (bx )dx =
1 , b
n > −1
a>b a
n > −1
J n (bx ) 1 dx = , x n
n = 1, 2, 3, ... e− b / 4a a 2
e − ax J 0 (b x ) dx =
xJ n (α x ) J n (β x )dx =
α J n (β ) J n′ (α ) − β J n (α ) J n′ (β ) β2 − α2
xJ n2 (α x )dx = 12 {J n′ (α )}2 + 12 (1 − n 2 /α 2 ){J n (α )}2 xJ 0 (α x )I 0 (β x )dx =
β J 0 (α )I 0′ (β ) − α J 0′ (α )I 0 (β ) α2 + β2
BESSEL FUNCTIONS
161
Integral Representations for Bessel Functions π
27.98.
J0 ( x ) =
1 π
∫
27.99.
Jn ( x ) =
1 π
∫
27.100.
Jn ( x ) =
27.101.
Y0 ( x ) = −
27.102.
I 0 (x) =
2
cos( x sin θ )dθ
0
π
0
xn π Γ (n + 12 )
n
2 π
1 π
cos(nθ − x sin θ )dθ
∫ π
π
0
cos( x sin θ ) cos 2 n θ dθ ,
cos( x cosh u)du
cosh ( x sin θ )dθ =
0
n > − 12
∞
0
∫
∫
n = integer
1 2π
∫
2π
0
e x sinθ dθ
Asymptotic Expansions 27.103.
Jn ( x ) ~
2 nπ π ⎞ ⎛ cos ⎜ x − − 2 4 ⎟⎠ πx ⎝
where x is large
27.104.
Yn ( x ) ~
2 nπ π ⎞ ⎛ sin ⎜ x − − 2 4 ⎟⎠ πx ⎝
where x is large
27.105.
Jn ( x ) ~
1 ⎛ ex ⎞ ⎜ ⎟ 2π n ⎝ 2n⎠
27.106.
Yn ( x ) ~ −
27.107
I n (x) ~
ex 2π x
where x is large
27.108
K n (x) ~
e− x 2π x
where x is large
2 ⎛ ex ⎞ π n ⎜⎝ 2n⎟⎠
n
where n is large
−n
where n is large
Orthogonal Series of Bessel Functions Let λ1 , λ2 , λ3 , ... be the positive roots of RJ n ( x ) + SxJ n′ ( x ) = 0, n > −1. Then the following series expansions hold under the conditions indicated. S = 0, R ≠ 0, i.e., λ1 , λ2 , λ3 , . . . are positive roots of JN(x) = 0 27.109.
f ( x ) = A1 J n (λ1 x ) + A2 J n (λ2 x ) + A3 J n (λ3 x ) + ⋅⋅⋅
where 27.110.
Ak =
2 J (λ k ) 2 n +1
∫
1
0
xf ( x ) J n (λ k x )dx
In psarticular if n = 0, 27.111.
f ( x ) = A1J 0 (λ1x ) + A2 J 0 (λ2 x ) + A3 J 0 (λ3 x ) + ⋅⋅⋅
where 27.112.
Ak =
2 J12 (λ k )
∫
1
0
xf ( x ) J 0 (λ k x )dx
BESSEL FUNCTIONS
162
R/S > −n 27.113.
f ( x ) = A1J n (λ1x ) + A2 J n (λ2 x ) + A3 J n (λ3 x ) + ⋅⋅⋅
where 27.114.
Ak =
2 J (λ k ) − J n−1 (λ k ) J n+1 (λ k ) 2 n
∫
1
xf ( x ) J n (λ k x )dx
0
In particular if n = 0. 27.115.
f ( x ) = A1J 0 (λ1x ) + A2 J 0 (λ2 x ) + A3 J 0 (λ3 x ) + ⋅⋅⋅
where 27.116.
Ak =
2 J 02 (λ k ) + J12 (λ k )
∫
1
0
xf ( x ) J 0 (λ k x )dx
The next formulas refer to the expansion of Bessel functions where S ≠ 0. R/S = −n 27.117.
f ( x ) = A0 x n + A1J n (λ1x ) + A2 J n (λ2 x ) + ⋅⋅⋅
where
27.118.
⎧A = 2(n + 1) 1 x n+1 f ( x )dx ∫0 ⎪⎪ 0 ⎨ 2 ⎪A = 2 ⎪⎩ k J n (λ k ) − J n−1 (λ k ) J n+1 (λ k )
1
∫
0
xf ( x ) J n (λ k x )dx
In particular if n = 0 so that R = 0 [i.e., l1, l2, l3, … are the positive roots of J1 (x) = 0], 27.119.
f ( x ) = A0 + A1J 0 (λ1x ) + A2 J 0 (λ2 x ) + ⋅⋅⋅
where 27.120.
⎧A = 2 1 xf ( x )dx ∫0 ⎪⎪ 0 ⎨ 1 ⎪A = 2 2 xf ( x ) J 0 (λ k x )dx k ∫ J 0 (λ k ) 0 ⎪⎩
R/S < −N In this case there are two pure imaginary roots ±il0 as well as the positive roots l1, l2, l3, … and we have 27.121.
f ( x ) = A0 I n (λ0 x ) + A1J n (λ1x ) + A2 J n (λ2 x ) + ⋅⋅⋅
where
27.122.
2 ⎧ ⎪A0 = I n2 (λ0 ) + I n−1 (λ0 )I n+1 (λ0 ) ⎪ ⎨ 2 ⎪A = ⎪⎩ k J n2 (λ k ) − J n−1 (λ k ) J n+1 (λ k )
∫
1
0
∫
1
0
xf ( x )I n (λ0 x )dx xf ( x ) J n (λ k x )dx
BESSEL FUNCTIONS
163
Miscellaneous Results 27.123.
cos ( x sin θ ) = J 0 ( x ) + 2 J 2 ( x ) cos 2θ + 2 J 4 ( x ) cos 4θ + ⋅⋅ ⋅
27.124.
sin ( x sin θ ) = 2 J1 ( x ) sin θ + 2 J3 ( x ) sin 3θ + 2 J5 ( x ) sin 5θ + ⋅⋅⋅
27.125.
J n ( x + y) =
∞
∑ J ( x)J
k =−∞
k
n−k
n = 0, ± 1, ± 2, ...
( y)
This is called the addition formula for Bessel functions. 27.126.
1 = J 0 ( x ) + 2 J 2 ( x ) + ⋅⋅⋅ + 2 J 2 n ( x ) + ⋅⋅⋅
27.127.
x = 2{J1 ( x ) + 3J3 ( x ) + 5J5 ( x ) + ⋅⋅⋅ + (2n + 1) J 2 n+1 ( x ) + ⋅ ⋅⋅}
27.128.
x 2 = 2{4 J 2 ( x ) + 16 J 4 ( x ) + 36 J6 ( x ) + ⋅⋅⋅ + (2n)2 J 2 n ( x ) + ⋅⋅⋅}
27.129.
xJ1 ( x ) = J 2 ( x ) − 2 J 4 ( x ) + 3J6 ( x ) − ⋅⋅⋅ 4
27.130.
1 = J 02 ( x ) + 2 J12 ( x ) + 2 J 22 ( x ) + 2 J32 ( x ) + ⋅⋅⋅
27.131.
J n′′( x ) = 14 {J n− 2 ( x ) − 2 J n ( x ) + J n+ 2 ( x )}
27.132.
J n′′′( x ) = 81 {J n −3 ( x ) − 3J n −1 ( x ) + 3J n+1 ( x ) − J n+3 ( x )}
Formulas 27.131 and 27.132 can be generalized.
2 sin nπ πx
27.133.
J n′ ( x ) J − n ( x ) − J −′ n J n ( x ) =
27.134.
J n ( x ) J − n+1 ( x ) + J − n ( x ) J n−1 ( x ) =
27.135.
J n+1 ( x )Yn ( x ) − J n ( x )Yn+1 ( x ) = J n ( x )Yn′( x ) − J n′ ( x )Yn ( x ) =
27.136.
sin x = 2{J1 ( x ) − J3 ( x ) + J5 ( x ) − ⋅⋅⋅}
27.137.
cos x = J 0 ( x ) − 2 J 2 ( x ) + 2 J 4 ( x ) − ⋅⋅⋅
27.138.
sinh x = 2{I1 ( x ) + I 3 ( x ) + I 5 ( x ) + ⋅⋅⋅}
27.139.
cosh x = I 0 ( x ) + 2{I 2 ( x ) + I 4 ( x ) + I 6 ( x ) + ⋅⋅⋅}
2 sin nπ πx 2 πx
LEGENDRE and ASSOCIATED LEGENDRE FUNCTIONS
28
Legendre’s Differential Equation 28.1.
(1 − x 2 ) y′′ − 2 xy′ + n(n + 1) y = 0
Solutions of this equation are called Legendre functions of order n.
Legendre Polynomials If n = 0, 1, 2, …, a solution of 28.1 is the Legendre polynomial Pn(x) given by Rodrigues’ formula 28.2.
Pn ( x ) =
1 dn 2 ( x − 1)n 2n n! dx n
Special Legendre Polynomials 28.3.
P0 ( x ) = 1
28.7.
P4 ( x ) = 81 (35x 4 − 30 x 2 + 3)
28.4.
P1 ( x ) = x
28.8.
P5 ( x ) = 81 (63x 5 − 70 x 3 + 15x )
28.5.
P2 ( x ) = 12 (3x 2 − 1)
28.9.
P6 ( x ) = 161 (231x 6 − 315x 4 + 105x 2 − 5)
28.6.
P3 ( x ) = 12 (5x 3 − 3x )
28.10.
P7 ( x ) = 161 (429 x 7 − 693x 5 + 315x 3 − 35x )
Legendre Polynomials in Terms of U where x ⴝ cos U 28.11.
P0 (cos θ ) = 1
28.12.
P1 (cos θ ) = cos θ
28.13.
P2 (cos θ ) = 14 (1 + 3 cos 2θ )
28.14.
P3 (cos θ ) = 81 (3 cos θ + 5 cos 3θ )
28.15.
P4 (cos θ ) =
164
1 64
(9 + 20 cos 2θ + 35 cos 4θ )
LEGENDRE AND ASSOCIATED LEGENDRE FUNCTIONS
28.16.
1 P5 (cos θ ) = 128 (30 cos θ + 35 cos 3θ + 63 cos 5θ )
28.17.
P6 (cos θ ) =
28.18.
1 P7 (cos θ ) = 1024 (175 cos θ + 189 cos 3θ + 231 cos 5θ + 429 cos 7θ )
1 512
(50 + 105 cos 2θ + 126 cos 4θ + 231 cos 6θ )
Generating Function for Legendre Polynomials ∞ 1 = Pn ( x )t n ∑ 1 − 2tx + t 2 n= 0
28.19.
Recurrence Formulas for Legendre Polynomials 28.20.
(n + 1)Pn+1 ( x ) − (2n + 1) x Pn ( x ) + nPn−1 ( x ) = 0
28.21.
Pn′+1 ( x ) − xPn′( x ) = (n + 1)Pn ( x )
28.22.
xPn′( x ) − Pn′−1 ( x ) = nPn ( x )
28.23.
Pn′+1 ( x ) − Pn′−1 ( x ) = (2n + 1)Pn ( x )
28.24.
( x 2 − 1)Pn′( x ) − nxPn ( x ) − nPn−1 ( x )
Orthogonality of Legendre Polynomials 28.25.
∫
1
28.26.
∫
1
−1
−1
Pm ( x )Pn ( x )dx = 0 m ≠ n {Pn ( x )}2 dx =
2 2n + 1
Because of 28.25, Pm(x) and Pn(x) are called orthogonal in –1 x 1.
Orthogonal Series of Legendre Polynomials 28.27.
f ( x ) = A0 P0 ( x ) + A1P1 ( x ) + A2 P2 ( x ) +
where 28.28.
Ak =
2k + 1 1 f ( x )Pk ( x )dx 2 ∫−1
165
LEGENDRE AND ASSOCIATED LEGENDRE FUNCTIONS
166
Special Results Involving Legendre Polynomials 28.29.
Pn (1) = 1
28.30.
Pn (−1) = (−1)n
28.31.
Pn (− x ) = (−1)n Pn ( x )
28.32.
⎧0 ⎪ Pn (0) = ⎨ 1 i 3 i 5 (n − 1) ⎪(−1)n / 2 2 i 4 i 6 n ⎩
28.33.
Pn ( x ) =
28.34.
∫ P ( x )dx =
28.35.
| Pn ( x ) | 1
28.36.
Pn ( x ) =
1 π
π
∫ (x + 0
1 2
n even
x 2 − 1 cos φ) dφ n
Pn+1 ( x ) − Pn−1 ( x ) 2n + 1
n
n +1
n odd
(z 2 − 1)n
π i ∫ (z − x ) c
n +1
dz
where C is a simple closed curve having x as interior point.
General Solution of Legendre’s Equation The general solution of Legendre’s equation is 28.37.
y = AU n ( x ) + BVn ( x )
where 28.38.
Un (x) = 1 −
n(n + 1) 2 n(n − 2)(n + 1)(n + 3) 4 x + x − 2! 4!
28.39.
Vn ( x ) = x −
(n − 1)(n + 2) 3 (n − 1)(n − 3)(n + 2)(n + 4) 5 x + x − 3! 5!
These series converge for –1 < x < 1.
Legendre Functions of the Second Kind If n = 0, 1, 2, … one of the series 28.38, 28.39 terminates. In such cases, 28.40.
⎧⎪U n ( x ) /U n (1) Pn ( x ) = ⎨ ⎪⎩Vn ( x ) /Vn (1)
n = 0, 2, 4, … n = 1, 3, 5, …
where 28.41.
U n (1) = (−1)
n/2
⎡⎛ n ⎞ ⎤ 2 ⎢⎜ ⎟ !⎥ ⎣⎝ 2⎠ ⎦ n
2
n!
n = 0, 2, 4, …
LEGENDRE AND ASSOCIATED LEGENDRE FUNCTIONS
28.42.
Vn (1) = (−1)
( n −1) / 2
2
n −1
⎡⎛ n − 1⎞ ⎤ ⎢⎜ ⎟ !⎥ ⎣⎝ 2 ⎠ ⎦
167
2
n!
n = 1, 3, 5,…
The nonterminating series in such a case with a suitable multiplicative constant is denoted by Qn(x) and is called Legendre’s function of the second kind of order n. We define
28.43.
⎧⎪U n (1)Vn ( x ) Qn ( x ) = ⎨ ⎩⎪−Vn (1)U n ( x )
n = 0, 2, 4,… n = 1, 3, 5,…
Special Legendre Functions of the Second Kind 28.44.
Q0 ( x ) =
1 ⎛1 + x⎞ ln 2 ⎜⎝ 1 − x ⎟⎠
28.45.
Q1 ( x ) =
x ⎛1 + x⎞ ln −1 2 ⎜⎝ 1 − x ⎟⎠
28.46.
Q2 ( x ) =
3x 2 − 1 ⎛ 1 + x ⎞ 3x ln ⎜ − 4 ⎝ 1 − x ⎟⎠ 2
28.47.
Q3 ( x ) =
5x 3 − 3x ⎛ 1 + x ⎞ 5x 2 2 ln ⎜ − + 4 2 3 ⎝ 1 − x ⎟⎠
The functions Qn(x) satisfy recurrence formulas exactly analogous to 28.20 through 28.24. Using these, the general solution of Legendre’s equation can also be written as 28.48.
y = APn ( x ) + BQn ( x )
Legendre’s Associated Differential Equation 28.49.
m2 ⎫ ⎧ (1 − x 2 ) y′′ − 2 xy′ + ⎨n(n + 1) − y=0 1 − x 2 ⎬⎭ ⎩
Solutions of this equation are called associated Legendre functions. We restrict ourselves to the important case where m, n are nonnegative integers.
Associated Legendre Functions of the First Kind 28.50.
Pnm ( x ) = (1 − x 2 )m / 2
dm (1 − x 2 )m / 2 d m + n 2 P ( x ) = ( x − 1)n dx m n 2n n! dx m + n
where Pn(x) are Legendre polynomials (page 164). We have 28.51.
Pn0 ( x ) = Pn ( x )
28.52.
Pnm ( x ) = 0
if m > n
LEGENDRE AND ASSOCIATED LEGENDRE FUNCTIONS
168
Special Associated Legendre Functions of the First Kind 28.53.
P11 ( x ) = (1 − x 2 )1/ 2
28.56.
P31 ( x ) = 32 (5x 2 − 1)(1 − x 2 )1/ 2
28.54.
P21 ( x ) = 3x (1 − x 2 )1/ 2
28.57.
P32 ( x ) = 15x (1 − x 2 )
28.55.
P22 ( x ) = 3(1 − x 2 )
28.58.
P33 ( x ) = 15(1 − x 2 )3 / 2
Generating Function for Pnm ( x) 28.59.
∞ (2m)!(1 − x 2 )m / 2 t m = Pnm ( x )t n 2m m !(1 − 2tx + t 2 )m +1/ 2 n∑ =m
Recurrence Formulas 28.60.
(n + 1 − m)Pnm+1 ( x ) − (2n + 1) x Pnm ( x ) + (n + m) Pnm−1 ( x ) = 0
28.61.
Pnm + 2 ( x ) −
2(m + 1) x m +1 P ( x ) + (n − m)(n + m + 1)Pnm ( x ) = 0 (1 − x 2 )1/ 2 n
Orthogonality of Pnm ( x) 1
28.62.
∫
28.63.
∫ {P
−1
Pl m ( x )P1m ( x )dx = 0
1
−1
m n
( x )} dx = 2
if n ≠ l
2 (n + m)! 2n + 1 (n − m)!
Orthogonal Series 28.64.
f ( x ) = Am Pmm ( x ) + Am +1Pmm+1 ( x ) + Am + 2 Pmm+ 2 ( x ) +
where 28.65.
Ak =
2k + 1 (k − m)! 1 f ( x )Pkm ( x )dx 2 (k + m)! ∫−1
Associated Legendre Functions of the Second Kind 28.66.
Qnm ( x ) = (1 − x 2 )m / 2
dm Q (x) dx m n
where Qn(x) are Legendre functions of the second kind (page 166). These functions are unbounded at x = ±1, whereas Pnm ( x ) are bounded at x = ± 1. The functions Qnm ( x ) satisfy the same recurrence relations as Pnm ( x ) (see 28.60 and 28.61).
General Solution of Legendre’s Associated Equation 28.67.
y = APnm ( x ) + BQnm ( x )
29
HERMITE POLYNOMIALS
Hermite’s Differential Equation 29.1.
y′′ − 2 xy′ + 2ny = 0
Hermite Polynomials If n = 0, 1, 2, …, then a solution of Hermite’s equation is the Hermite polynomial Hn(x) given by Rodrigue’s formula. 29.2.
H n ( x ) = (−1)n e x
2
d n −x (e ) dx n 2
Special Hermite Polynomials 29.3.
H0 (x) = 1
29.7.
H 4 ( x ) = 16 x 4 − 48 x 2 + 12
29.4.
H1 ( x ) = 2 x
29.8.
H 5 ( x ) = 32 x 5 − 160 x 3 + 120 x
29.5.
H2 (x) = 4 x 2 − 2
29.9.
H 6 ( x ) = 64 x 6 − 480 x 4 + 720 x 2 − 120
29.6.
H 3 ( x ) = 8 x 3 − 12 x
29.10.
H 7 ( x ) = 128 x 7 − 1344 x 5 + 3360 x 3 − 1680 x
Generating Function ∞
29.11.
e 2 tx − t = ∑ 2
n= 0
H n ( x )t n n!
Recurrence Formulas 29.12.
H n+1 ( x ) = 2 xH n ( x ) − 2nH n−1 ( x )
29.13.
H n′ ( x ) = 2nH n−1 ( x )
Orthogonality of Hermite Polynomials 29.14.
∫
29.15.
∫
∞ −∞ ∞
−∞
e − x H m ( x ) H n ( x )dx = 0 2
m≠n
e − x {H n ( x )}2 dx = 2n n! π 2
169
HERMITE POLYNOMIALS
170
Orthogonal Series 29.16.
f ( x ) = A0 H 0 ( x ) + A1H1 ( x ) + A2 H 2 ( x ) +
where 29.17.
Ak =
1 2 k! π
∫
∞
−∞
k
e − x f ( x ) H k ( x )dx 2
Special Results n(n − 1) n(n − 1)(n − 2)(n − 3) (2 x )n− 4 − (2 x ) n − 2 + 2! 1!
29.18.
H n ( x ) = (2 x ) n −
29.19.
H n (− x ) = (−1)n H n ( x )
29.20.
H 2 n−1 (0) = 0
29.21.
H 2 n (0) = (−1)n 2n i 1 i 3 i 5 (2n − 1)
29.22.
∫
29.23.
d −x {e H n ( x )} = −e − x H n+1 ( x ) dx
29.24.
∫
x
29.25.
∫
∞
29.26.
H n ( x + y) = ∑
x
H n (t )dt =
0
H n+1 ( x ) H n+1 (0) − 2(n + 1) 2(n + 1)
2
0
2
e − t H n (t )dt = H n−1 (0) − e − x H n−1 ( x )
−∞
2
2
t ne − t H n ( xt )dt = π n! Pn ( x ) 2
1 ⎛ n⎞ n / 2 ⎜ ⎟ H k ( x 2 ) H n− k ( y 2 ) 2 ⎝ k⎠ k =0 n
This is called the addition formula for Hermite polynomials. n
29.27.
∑ k =0
H k ( x ) H k ( y) H n+1 ( x ) H n ( y) − H n ( x ) H n+1 ( y) = 2k k ! 2n+11 n!( x − y)
30
LAGUERRE and ASSOCIATED LAGUERRE POLYNOMIALS
Laguerre’s Differential Equation 30.1.
xy ′′ + (1 − x ) y ′ + ny = 0
Laguerre Polynomials If n = 0, 1, 2, …, then a solution of Laguerre’s equation is the Laguerre polynomial Ln(x) given by Rodrigues’ formula 30.2.
Ln ( x ) = e x
d n n −x (x e ) dx n
Special Laguerre Polynomials 30.3.
L0 ( x ) = 1
30.4.
L1 ( x ) = − x + 1
30.5.
L2 ( x ) = x 2 − 4 x + 2
30.6.
L3 ( x ) = − x 3 + 9 x 2 − 18 x + 6
30.7.
L4 ( x ) = x 4 − 16 x 3 + 72 x 2 − 96 x + 24
30.8.
L5 ( x ) = − x 5 + 25x 4 − 200 x 3 + 600 x 2 − 600 x + 120
30.9.
L6 ( x ) = x 6 − 36 x 5 + 450 x 4 − 2400 x 3 + 5400 x 2 − 4320 x + 720
30.10.
L7 ( x ) = − x 7 + 49 x 6 − 882 x 5 + 7350 x 4 − 29, 400 x 3 + 52, 920 x 2 − 35, 280 x + 5040
Generating Function 30.11.
∞ L ( x )t n e − xt /(1− t ) =∑ n n! 1− t n= 0
171
LAGUERRE AND ASSOCIATED LAGUERRE POLYNOMIALS
172
Recurrence Formulas 30.12.
Ln+1 ( x ) − (2n + 1 − x ) Ln ( x ) + n 2 Ln−1 ( x ) = 0
30.13.
Ln′ ( x ) − nLn′−1 ( x ) + nLn −1 ( x ) = 0
30.14.
xLn′ ( x ) = nLn ( x ) − n 2 Ln−1 ( x )
Orthogonality of Laguerre Polynomials 30.15.
∫
∞
30.16.
∫
∞
0
0
e − x Lm ( x ) Ln ( x )dx = 0
m≠n
e − x {Ln ( x )}2 dx = (n!)2
Orthogonal Series 30.17.
f ( x ) = A0 L0 ( x ) + A1L1 ( x ) + A2 L2 ( x ) +
where 30.18.
Ak =
1 (k !)2
∫
∞
0
e − x f ( x ) Lk ( x )dx
Special Results 30.19.
Ln ( 0 ) = n !
30.20.
∫
30.21.
n 2 x n−1 n 2 (n − 1)2 x n− 2 ⎫ ⎧ Ln ( x ) = (−1)n ⎨x n − + − (−1)n n!⎬ 1 ! 2 ! ⎭ ⎩
30.22.
⎧⎪ 0 p −x ( ) x e L x dx = ⎨ ∫0 n 2 n ⎩⎪(−1) (n !)
30.23.
∑
x
0
Ln (t )dt = Ln ( x ) −
Ln+1 ( x ) n +1
∞
n
k =0
if p < n if p = n
Lk ( x ) Lk ( y) Ln ( x ) Ln+1 ( y) − Ln+1 ( x ) Ln ( y) = (k !)2 (n!)2 ( x − y)
∞
t k Lk ( x ) = e t J 0 (2 xt ) (k !)2 k =0
30.24.
∑
30.25.
Ln ( x ) = ∫ u ne x −u J 0 (2 xu )du
∞
0
LAGUERRE AND ASSOCIATED LAGUERRE POLYNOMIALS
173
Laguerre’s Associated Differential Equation 30.26.
xy′′ + (m + 1 − x ) y′ + (n − m) y = 0
Associated Laguerre Polynomials Solutions of 30.26 for nonnegative integers m and n are given by the associated Laguerre polynomials 30.27.
Lmn ( x ) =
dm L (x) dx m n
where Ln(x) are Laguerre polynomials (see page 171). 30.28.
L0n ( x ) = Ln ( x )
30.29.
Lmn ( x ) = 0
if m > n
Special Associated Laguerre Polynomials 30.30.
L11 ( x ) = −1
30.35.
L33 ( x ) = −6
30.31.
L12 ( x ) = 2 x − 4
30.36.
L14 ( x ) = 4 x 3 − 48 x 2 + 144 x − 96
30.32.
L22 ( x ) = 2
30.37.
L24 ( x ) = 12 x 2 − 96 x + 144
30.33.
L13 ( x ) = −3x 2 + 18 x − 18
30.38.
L34 ( x ) = 24 x − 96
30.34.
L23 ( x ) = −6 x + 18
30.39.
L44 ( x ) = 24
Generating Function for Lmn ( x) 30.40.
∞ Lmn ( x ) n (−1)m t m − xt /(1− t ) e = t ∑ n! (1 − t )m+1 n= m
Recurrence Formulas 30.41.
n − m +1 m Ln+1 ( x ) + ( x + m − 2n − 1) Lmn ( x ) + n 2 Lmn−1 ( x ) = 0 n +1
30.42.
d m {L ( x )} = Lmn +1 ( x ) dx n
30.43.
d m −x m {x e Ln ( x )} = (m − n − 1) x m −1e − x Lmn −1 ( x ) dx
30.44.
x
d m {L ( x )} = ( x − m) Lmn ( x ) + (m − n − 1) Lmn −1 ( x ) dx n
LAGUERRE AND ASSOCIATED LAGUERRE POLYNOMIALS
174
Orthogonality 30.45.
∫
∞
30.46.
∫
∞
0
0
x me − x Lmn ( x ) Lmp ( x )dx = 0 x me − x {Lmn ( x )}2 dx =
p≠n
(n!)3 (n − m)!
Orthogonal Series 30.47.
f ( x ) = Am Lmm ( x ) + Am +1Lmm +1 ( x ) + Am + 2 Lmm+ 2 ( x ) +
where 30.48.
Ak =
(k − m)! ∞ m − x m x e Lk ( x ) f ( x )dx (k !)3 ∫0
Special Results 30.49. 30.50.
Lmn ( x ) = (−1)n
∫
∞
0
{
}
n! n(n − m) n− m −1 n(n − 1)(n − m)(n − m − 1) n− m − 2 x + x n−m − x + 2! (n − m)! 1!
x m +1e − x {Lmn ( x )}2 dx =
(2n − m + 1)(n!)3 (n − m)!
31
CHEBYSHEV POLYNOMIALS
Chebyshev’s Differential Equation 31.1.
(1 − x 2 ) y n − xy ′ + n 2 y = 0
n = 0, 1, 2, ...
Chebyshev Polynomials of the First Kind A solution of 31.1 is given by 31.2.
⎛ n⎞ ⎛ n⎞ Tn ( x ) = cos (n cos −1 x ) = x n − ⎜ ⎟ x n − 2 (1 − x 2 ) + ⎜ ⎟ x n − 4 (1 − x 2 )2 − ⎝ 4⎠ ⎝ 2⎠
Special Chebyshev Polynomials of The First Kind 31.3.
T0 ( x ) = 1
31.7.
T4 ( x ) = 8 x 4 − 8 x 2 + 1
31.4.
T1 ( x ) = x
31.8.
T5 ( x ) = 16 x 5 − 20 x 3 + 5x
31.5.
T2 ( x ) = 2 x 2 − 1
31.9.
T6 ( x ) = 32 x 6 − 48 x 4 + 18 x 2 − 1
31.6.
T3 ( x ) = 4 x 3 − 3x
31.10.
T7 ( x ) = 64 x 7 − 112 x 5 + 56 x 3 − 7 x
Generating Function for Tn ( x) 31.11.
∞ 1 − tx n 2 = ∑ Tn ( x )t 1 − 2tx + t n=0
Special Values 31.12.
Tn (− x ) = (−1)n Tn ( x )
31.14.
Tn (−1) = (−1)n
31.13.
Tn (1) = 1
31.15.
T2 n (0) = (−1)n
31.16.
T2 n+1 (0) = 0
Recursion Formula for Tn ( x) 31.17.
Tn+1 ( x ) − 2 xTn ( x ) + Tn−1 ( x ) = 0
175
CHEBYSHEV POLYNOMIALS
176
Orthogonality 31.18.
31.19.
∫
1
∫
1
−1
Tm ( x )Tn ( x ) dx = 0 1− x2 ⎧ π dx = ⎨ 1− x ⎩π / 2
{Tn ( x )}2
−1
2
m≠n if n = 0 if n = 1, 2, ...
Orthogonal Series 31.20.
f ( x ) = 12 A0T0 ( x ) + A1T1 ( x ) + A2T2 ( x ) +
where 31.21.
Ak =
2 π
∫
1
−1
f ( x )Tk ( x ) dx 1− x2
Chebyshev Polynomials of The Second Kind 31.22.
Un (x) =
sin{(n + 1) cos −1 x} sin (cos −1 x )
⎛ n + 1⎞ n ⎛ n + 1⎞ n − 2 ⎛ n + 1⎞ n − 4 x −⎜ x (1 − x 2 ) + ⎜ x (1 − x 2 )2 − =⎜ ⎝ 3 ⎟⎠ ⎝ 1 ⎠⎟ ⎝ 5 ⎟⎠
Special Chebyshev Polynomials of The Second Kind 31.23.
U0 (x) = 1
31.27.
U 4 ( x ) = 16 x 4 − 12 x 2 + 1
31.24.
U1 ( x ) = 2 x
31.28.
U 5 ( x ) = 32 x 5 − 32 x 3 + 6 x
31.25.
U2 (x) = 4 x 2 − 1
31.29.
U 6 ( x ) = 64 x 6 − 80 x 4 + 24 x 2 − 1
31.26.
U3 ( x) = 8 x 3 − 4 x
31.30.
U 7 ( x ) = 128 x 7 − 192 x 5 + 80 x 3 − 8 x
Generating Function for Un ( x) 31.31.
∞ 1 n 2 = ∑ U n ( x )t 1 − 2tx + t n= 0
Special Values 31.32.
U n (− x ) = (−1)nU n ( x )
31.34.
U n (−1) = (−1)n (n + 1)
31.33.
U n (1) = n + 1
31.35.
U 2 n (0) = (−1)n
31.36.
U 2 n+1 (0) = 0
CHEBYSHEV POLYNOMIALS
177
Recursion Formula for Un ( x) 31.37.
U n+1 ( x ) − 2 xU n ( x ) + U n−1 ( x ) = 0
Orthogonality 31.38.
∫
31.39.
∫
1
1 − x 2 U m ( x )U n ( x )dx = 0
−1 1
−1
1 − x 2 {U n ( x )}2 dx =
m≠n
π 2
Orthogonal Series 31.40.
f ( x ) = A0U 0 ( x ) + AU 1 1 ( x ) + A2U 2 ( x ) +
where 31.41.
Ak =
2 π
∫
1
1 − x 2 f ( x )U k ( x )dx
−1
Relationships Between Tn ( x) and Un ( x) 31.42.
Tn ( x ) = U n ( x ) − xU n−1 ( x )
31.43.
(1 − x 2 )U n−1 ( x ) = xTn ( x ) − Tn+1 ( x )
31.44.
Un (x) =
1 π
∫
31.45.
Tn ( x ) =
1 π
∫
1
−1
1
−1
Tn+1 ( )d ( − x ) 1 − 2 1 − 2 U n−1 ( ) d x−
General Solution of Chebyshev’s Differential Equation 31.46.
⎧AT ( x ) + B 1 − x 2 U ( x ) ⎪ n n −1 y=⎨ −1 ⎩⎪A + B sin x
if n = 1, 2, 3, … if n = 0
32
HYPERGEOMETRIC FUNCTIONS
Hypergeometric Differential Equation 32.1.
x (1 − x ) y n + {c − (a + b + 1) x}y′ − aby = 0
Hypergeometric Functions A solution of 32.1 is given by 32.2.
F (a, b; c; x ) = 1 +
aib a(a + 1)b(b + 1) 2 a(a + 1)(a + 2)b(b + 1)(b + 2) 3 x + x+ x + 1i 2 i 3 i c(c + 1)(c + 2) 1i c 1i 2 i c(c + 1)
If a, b, c are real, then the series converges for –1 < x < 1 provided that c – (a + b) > –1.
Special Cases 32.3.
F (− p,1;1; − x ) = (1 + x ) p
32.8.
F ( 12 , 12 ; 32 ; x 2 ) = (sin −1 x ) /x
32.4.
F (1,1; 2; − x ) = [ln (1 + x )]/x
32.9.
F ( 12 ,1; 32 ; − x 2 ) = (tan −1 x ) /x
32.5.
lim F (1, n;1; x /n) = e x
32.10.
F (1, p; p; x ) = 1/ (1 − x )
32.6.
F ( 12 , − 12 ; 12 ; sin 2 x ) = cos x
32.11.
F (n + 1, − n;1;(1 − x ) / 2) = Pn ( x )
32.7.
F ( 12 ,1;1; sin 2 x ) = sec x
32.12.
F (n, − n; 12 ;(1 − x ) / 2) = Tn ( x )
n→∞
General Solution of The Hypergeometric Equation If c, a – b and c – a – b are all nonintegers, then the general solution valid for | x | < 1 is 32.13.
178
y = AF (a, b; c; x ) + Bx 1−c F (a − c + 1, b − c + 1; 2 − c; x )
HYPERGEOMETRIC FUNCTIONS
Miscellaneous Properties Γ (c)Γ (c − a − b) Γ (c − a)Γ (c − b)
32.14.
F (a, b; c;1) =
32.15.
d ab F (a, b; c; x ) = F (a + 1, b + 1; c + 1; x ) dx c
32.16.
F (a, b; c; x ) =
32.17.
F (a, b; c; x ) = (1 − x )c− a− b F (c − a, c − b; c; x )
1 Γ (c) u b−1 (1 − u)c− b−1 (1 − ux )− a du ∫ Γ (b)Γ (c − b) 0
179
Section VIII: Laplace and Fourier Transforms
33
LAPLACE TRANSFORMS
Definition of the Laplace Transform of F(t) 33.1.
{F (t )} =
∫
∞ 0
e − st F (t )dt = f (s)
In general f (s) will exist for s > a where a is some constant. is called the Laplace transform operator.
Definition of the Inverse Laplace Transform of f(s) If {F(t)} = f (s), then we say that F (t) = –1{f (s)} is the inverse Laplace transform of f (s). –1 is called the inverse Laplace transform operator.
Complex Inversion Formula The inverse Laplace transform of f (s) can be found directly by methods of complex variable theory. The result is
33.2.
F (t ) =
1 2π i
∫
c + i∞ c − i∞
e st f (s)ds =
1 lim 2π i T →∞
∫
c + iT c − iT
e st f (s)ds
where c is chosen so that all the singular points of f (s) lie to the left of the line Re{s} = c in the complex s plane.
180 Copyright © 2009, 1999, 1968 by The McGraw-Hill Companies, Inc. Click here for terms of use.
LAPLACE TRANSFORMS
181
Table of General Properties of Laplace Transforms
f (s)
F(t)
33.3.
a f1 (s) + bf2 (s)
aF1 (t ) + bF2 (t )
33.4.
f (s /a)
a F (at )
33.5.
f (s – a)
eatF(t)
33.6.
e–asf (s)
33.7.
sf (s) – F (0)
F ′ (t )
33.8.
s 2 f (s) − sF (0) − F ′(0)
F ′′(t )
33.9.
s n f (s) − s n−1 F (0) − s n− 2 F ′(0) − − F ( n−1) (0)
F(n)(t)
33.10.
f ′ (s )
–tF(t)
33.11.
f ′′(s)
t2F(t)
33.12.
f (n)(s)
(–1)nt nF(t)
33.13.
f (s ) s
33.14.
f (s ) sn
33.15.
f (s)g(s)
(t − a) =
∫ ∫
t 0
t
{
F (t − a) t > a 0 t
t 0
F (u)du
∫ F (u)du n = 0
∫
t 0
(t − u)n−1 ∫0 (n − 1)! F (u)du t
F (u)G (t − u)du
LAPLACE TRANSFORMS
182
33.16.
∫
33.17.
1 1 − e − sT
F(t)
f (u)du
F (t ) t
∫
T 0
e − su F (u)du
F(t) = F(t + T)
1 πt
f ( s) s
33.18.
()
1 1 f s s
33.19.
1
33.20.
s
n+1
f
∫
(1s)
1 2 π
∫
∞ 0
u −3 / 2 e − s
2
/ 4u
∞ 0
∫
∞ 0
e−u
2
/ 4t
F (u)du
J 0 (2 ut )F (u)du
∞
t n / 2 ∫ u − n / 2 J n (2 ut )F (u)du 0
f (s + 1/s) s2 + 1
33.21.
33.22.
∞ s
f (s)
∫
t 0
J 0 (2 u(t − u)) F (u)du
f (u)du
F(t2)
∞
t u F (u) du Γ (u + 1)
33.23.
f (ln s) s ln s
∫
33.24.
P (s ) Q (s )
∑ Q′(α
P(s) = polynomial of degree less than n, Q(s) = (s – a1)(s – a2) … (s – an) where a1, a2, …, an are all distinct.
0
n
k =1
P(α k ) α t e k) k
LAPLACE TRANSFORMS
183
Table of Special Laplace Transforms f (s)
F(t)
33.25.
1 s
1
33.26.
1 s2
t
33.27.
33.28.
1 sn
n = 1, 2, 3,… 1 sn
1 s−a
33.29.
33.30.
33.31.
n>0
1 (s − a) n
t n−1 Γ(n)
eat
n = 1, 2, 3,…
1 (s − a) n
t n−1 , 0! = 1 (n − 1)!
n>0
t n−1e at , 0! = 1 (n − 1)! t n−1e at Γ(n)
33.32.
1 s + a2
sin at a
33.33.
s s2 + a2
cos at
33.34.
1 (s − b) 2 + a 2
e bt sin at a
33.35.
s−b (s − b) 2 + a 2
e bt cos at
33.36.
1 s2 − a2
sinh at a
33.37.
s s2 − a2
cosh at
33.38.
1 (s − b) 2 − a 2
e bt sinh at a
2
LAPLACE TRANSFORMS
184
33.39.
f (s)
F(t)
s−b (s − b) 2 − a 2
e bt cosh at
33.40.
1 (s − a)(s − b)
a≠b
e bt − e at b−a
33.41.
s (s − a)(s − b)
a≠b
be bt − ae at b−a
33.42.
1 (s 2 + a 2 ) 2
sin at − at cos at 2a 3
33.43.
s (s 2 + a 2 ) 2
t sin at 2a
33.44.
s2 (s + a 2 ) 2
sin at + at cos at 2a
33.45.
s3 (s + a 2 ) 2
cos at − 12 at sin at
33.46.
s2 − a2 (s 2 + a 2 ) 2
t cos at
33.47.
1 (s 2 − a 2 ) 2
at cosh at − sinh at 2a 3
33.48.
s (s 2 − a 2 ) 2
t sinh at 2a
33.49.
s2 (s 2 − a 2 ) 2
sinh at + at cosh at 2a
33.50.
s3 (s − a 2 ) 2
cosh at + 12 at sinh at
33.51.
s2 (s − a 2 )3 / 2
t cosh at
33.52.
1 (s + a 2 ) 3
(3 − a 2 t 2 ) sin at − 3at cos at 8a 5
33.53.
s (s + a 2 )3
t sin at − at 2 cos at 8a 3
33.54.
s2 (s + a 2 )3
(1 + a 2 t 2 ) sin at − at cos at 8a 3
33.55.
s3 (s 2 + a 2 )3
3t sin at + at 2 cos at 8a
2
2
2
2
2
2
2
LAPLACE TRANSFORMS
185
f (s)
F(t)
33.56.
s4 (s + a 2 )3
(3 − a 2 t 2 ) sin at + 5at cos at 8a
33.57.
s5 (s + a 2 )3
(8 − a 2 t 2 ) cos at − 7at sin at 8
33.58.
3s 2 − a 2 (s 2 + a 2 ) 3
t 2 sin at 2a
33.59.
s 3 − 3a 2 s (s 2 + a 2 )3
1 2
t 2 cos at
33.60.
s 4 − 6a 2 s 2 + a 4 (s 2 + a 2 ) 4
1 6
t 3 cos at
33.61.
s3 − a2s (s 2 + a 2 ) 4
t 3 sin at 24 a
33.62.
1 (s − a 2 ) 3
(3 + a 2 t 2 ) sinh at − 3at cosh at 8a 5
33.63.
s (s − a 2 ) 3
at 2 cosh at − t sinh at 8a 3
33.64.
s2 (s − a 2 )3
at cosh at + (a 2 t 2 − 1) sinh at 8a 3
33.65.
s3 (s − a 2 ) 3
3t sinh at + at 2 cosh at 8a
33.66.
s4 (s − a 2 )3
(3 + a 2 t 2 ) sinh at + 5at cosh at 8a
33.67.
s5 (s − a 2 )3
(8 + a 2 t 2 ) cosh at + 7at sinh at 8
33.68.
3s 2 + a 2 (s 2 − a 2 ) 3
t 2 sinh at 2a
33.69.
s 3 + 3a 2 s (s 2 − a 2 )3
1 2
t 2 cosh at
33.70.
s 4 + 6a 2 s 2 + a 4 (s 2 − a 2 ) 4
1 6
t 3 cosh at
33.71.
s3 + a2s (s 2 − a 2 ) 4
t 3 sinh at 24 a
33.72.
1 s + a3
⎫ e at / 2 ⎧ 3at 3at − cos + e −3at / 2 ⎬ 2 ⎨ 3 sin 2 2 3a ⎩ ⎭
2
2
2
2
2
2
2
2
3
LAPLACE TRANSFORMS
186
f (s)
F(t)
33.73.
s 3 s + a3
⎫ 3at e at / 2 ⎧ 3at cos + 3 sin − e −3at / 2 ⎬ 3a ⎨⎩ 2 2 ⎭
33.74.
s2 s3 + a3
1 ⎛ − at 3at ⎞ e + 2e at / 2 cos 3 ⎜⎝ 2 ⎟⎠
33.75.
1 s − a3
e − at / 2 3a 2
33.76.
s s − a3
⎫ e − at / 2 ⎧ 3at 3at − cos + e3at / 2 ⎬ 3 sin 3a ⎨⎩ 2 2 ⎭
33.77.
s2 s − a3
1 ⎛ at 3at ⎞ e + 2e − at / 2 cos 3 ⎜⎝ 2 ⎟⎠
33.78.
1 s 4 + 4a 4
1 (sin at cosh at − cos at sinh at ) 4a3
33.79.
s s 4 + 4a 4
sin at sinh at 2a 2
33.80.
s2 s + 4a 4
1 (sin at cosh at + cos at sinh at ) 2a
33.81.
s3 s + 4a 4
cos at cosh at
33.82.
1 s4 − a4
1 (sinh at − sin at ) 2a 3
33.83.
s s − a4
1 (cosh at − cos at ) 2a 2
33.84.
s2 s − a4
1 (sinh at + sin at ) 2a
33.85.
s3 s − a4
33.86.
33.87.
3
3
3
4
⎧ 3at / 2 3at 3at ⎫ − cos − 3 sin ⎨e 2 2 ⎬⎭ ⎩
4
4
4
4
1 s+a + s+b 1 s s+a
1 2
(cosh at + cos at ) e − bt − e − at 2(b − a) π t 3 erf at a
33.88.
1 s (s − a)
e at erf at a
33.89.
1 s−a +b
⎫ ⎧ 1 e at ⎨ − b e b t erfc(b t )⎬ ⎭ ⎩ πt 2
LAPLACE TRANSFORMS
187
f (s)
F(t)
33.90.
1 s2 + a2
J 0 (at )
33.91.
1 s2 − a2
I 0 (at )
33.92.
( s 2 + a 2 − s) n s2 + a2
n > −1
a n J n (at )
33.93.
(s − s 2 − a 2 ) n s2 − a2
n > −1
a n I n (at )
33.94.
e b(s− s +a ) s2 + a2
J 0 (a t (t + 2b))
33.95.
e− b s +a s2 + a2
⎧J 0 (a t 2 − b 2 ) t > b ⎨ t
33.96.
1 (s 2 + a 2 ) 3 / 2
tJ1 (at ) a
33.97.
s (s 2 + a 2 )3 / 2
tJ 0 (at )
33.98.
s2 (s + a 2 )3 / 2
J 0 (at ) − atJ1 (at )
33.99.
1 (s − a 2 )3 / 2
tI1 (at ) a
33.100.
s (s 2 − a 2 ) 3 / 2
tI 0 (at )
33.101.
s2 (s − a 2 )3 / 2
I 0 (at ) + atI1 (at )
1 e− s = s(e − 1) s(1 − e − s )
F (t ) = n, n t < n + 1, n = 0,1, 2,…
2
2
33.102.
2
2
2
2
2
s
See also entry 33.165. 33.103.
33.104.
1 e− s = s s(e − r ) s(1 − re − s ) es − 1 1 − e− s = s s(e − r ) s(1 − re − s )
[t ]
F (t ) = ∑ r k k =1
where [t] = greatest integer t F (t ) = r n , n t < n + 1, n = 0,1, 2,…
See also entry 33.167. 33.105.
e − a /s s
cos 2 at πt
LAPLACE TRANSFORMS
188
33.106.
33.107.
f (s)
F(t)
e − a /s s3/ 2
sin 2 at πa
e − a /s s n+1
n > −1
33.108.
e− a s s
33.109.
e− a
33.110.
1 − e− a s
33.111.
e− a s
33.113.
33.114.
e− a / s s n+1 ln
J n (2 at ) e− a / 4t πt 2
a e− a 2 π t3
s
2
/ 4t
s
erf (a / 2 t )
s
erfc(a / 2 t )
e− a s s ( s + b)
33.112.
n/2
⎛ t⎞ ⎝ a⎠
n > −1
⎛ s + a⎞ ⎝ s + b⎠
a ⎞ ⎛ e b ( bt + a ) erfc ⎜ b t + ⎝ 2 t ⎟⎠ 1 π ta 2 n+1
∫
∞ 0
une−u
2
/ 4 a2t
J 2 n (2 u )du
e − bt − e − at t
33.115.
ln[(s 2 + a 2 ) /a 2 ] 2s
Ci(at )
33.116.
ln[(s + a) /a] s
Ei(at )
33.117.
(γ + ln s) s γ = Euler’s constant = .5772156 …
ln t
33.118.
⎛ s2 + a2 ⎞ ln ⎜ 2 ⎝ s + b 2 ⎟⎠
2(cos at − cos bt ) t
33.119.
π 2 (γ + ln s)2 + 6s s γ = Euler’s constant = .5772156 …
ln 2 t
33.120.
ln s s
33.121.
ln 2 s s
−
−(ln t + γ )
γ = Euler’s constant = .5772156 … (ln t + γ )2 − 16 π 2
γ = Euler’s constant = .5772156 …
LAPLACE TRANSFORMS
189
f (s) 33.122.
F(t)
Γ ′(n + 1) − Γ (n + 1) ln s s n+1
n > −1
t n ln t
33.123.
tan −1 (a /s)
sin at t
33.124.
tan −1 (a /s) s
Si(at )
33.125.
e a /s erfc( a /s) s
e −2 at πt
33.126.
es
2
/ 4 a2
erfc(s / 2a)
2a − a t e π
es
2
/ 4 a2
erfc(s / 2a) s
erf(at )
33.127.
2 2
33.128.
e as erfc as s
1 π (t + a)
33.129.
e as Ei(as)
1 t+a
33.130.
1⎡ π ⎤ cos as − Si(as) − sin as Ci(as)⎥ a ⎢⎣ 2 ⎦
33.131.
33.132.
33.133.
sin as
{
}
{ {
} }
{
}
π − Si(as) + cos as Ci(as) 2
t t 2 + a2
π − Si(as) − sin as Ci(as) 2 s
tan −1 (t /a)
cos as
sin as
1 t 2 + a2
π − Si(as) − cos as Ci(as) 2 s
1 ⎛ t 2 + a2 ⎞ ln 2 ⎜⎝ a 2 ⎟⎠
33.134.
⎡π − Si(as)⎤ + Ci 2 (as) ⎢⎣ 2 ⎥⎦
1 ⎛ t 2 + a2 ⎞ ln t ⎜⎝ a 2 ⎟⎠
33.135.
0
(t) = null function
33.136.
1
δ (t) = delta function
33.137.
e − as
δ (t − a)
33.138.
e − as s See also entry 33.163.
(t − a)
2
LAPLACE TRANSFORMS
190
f (s)
F(t)
33.139.
sinh sx s sinh sa
x 2 ∞ (−1)n nπ x nπ t sin cos + a π∑ n a a n =1
33.140.
sinh sx s cosh sa
4 ∞ (−1)n (2n − 1)π x (2n − 1)π t sin sin 2 − 1 2 2a π∑ n a n =1
33.141.
cosh sx s sinh as
t 2 ∞ (−1)n nπ x nπ t cos sin + a π∑ n a a n =1
33.142.
cosh sx s cosh sa
33.143.
sinh sx s sinh sa
33.144.
sinh sx s cosh sa
33.145.
cosh sx s 2 sinh sa
33.146.
cosh sx s cosh sa
33.147.
cosh sx s 3 cosh sa
33.148.
33.149.
1+
(2n − 1)π x (2n − 1)π t 4 ∞ (−1)n cos cos 2 − 1 2 2a π∑ n a n =1 xt 2a ∞ (−1)n nπ x nπ t sin sin + 2 a π2 ∑ a a n n =1
2
2
2
x+
(2n − 1)π x (2n − 1)π t 8a ∞ (−1)n cos 2 sin 2 a 2a ( ) 2 n − 1 π2 ∑ n=1 t 2 2a ∞ (−1)n nπ x ⎛ nπ t ⎞ cos + 1 − cos 2 a a ⎠ 2a π 2 ∑ ⎝ n n =1
t+
8a ∞ (−1)n (2n − 1)π x (2n − 1)π t sin 2 cos 2 2a a ( 2 1 ) n − π2 ∑ n =1
2π a2
sinh x s sinh a s cosh x s cosh a s
∞
1 2 16a 2 (t + x 2 − a 2 ) − 3 2 π
π a2
∞
∑ (−1)
n −1
(−1)n
∑ (2n − 1)
3
cos
n=1
∞
∑ (−1)
n
(2n − 1)π x (2n − 1)π t cos 2a 2a
ne − n π t /a sin 2
2
2
n =1
nπ x a
(2n − 1)e − ( 2 n−1) π t / 4 a cos 2
2
2
n =1
(2n − 1)π x 2a
33.150.
sinh x s s cosh a s
2 ∞ (2n − 1)π x (−1)n−1 e − ( 2 n−1) π t / 4 a sin 2a a∑ n =1
33.151.
cosh x s s sinh a s
1 2 ∞ nπ x + (−1)n e − n π t /a cos a a∑ a n =1
33.152.
sinh x s s sinh a s
x 2 ∞ (−1)n − n π t /a nπ x sin e + ∑ a π n=1 n a
33.153.
cosh x s s cosh a s
33.154.
sinh x s 2 s sinh a s
33.155.
cosh x s 2 s cosh a s
2
2
2
2
2
2
1+
2
2
2
(2n − 1)π x 4 ∞ (−1)n − ( 2 n−1) π t / 4 a cos e 2a 2 − 1 π∑ n n =1
xt 2a 2 + a π3
2
2
2
(−1)n nπ x (1 − e − n π t /a ) sin 3 a n n =1 ∞
∑
1 2 16a 2 (x − a2 ) + t − 3 2 π
∞
2
(−1)n
∑ (2n − 1) n =1
2
2
e − ( 2 n−1) π t / 4 a cos 2
3
2
2
(2n − 1)π x 2a
LAPLACE TRANSFORMS
191
f (s)
F(t) e − λ t /a J 0 (λn x /a) λn J1 (λn ) n =1 ∞
J 0 (ix s ) s J 0 (ia s )
33.156.
1 − 2∑
2 n
2
where λl, λ2,… are the positive roots of J0(λ) = 0 ∞ e − λ t /a J 0 (λn x /a) 1 2 ( x − a 2 ) + t + 2a 2 ∑ 4 λn3 J1 (λn ) n =1 2 n
33.157.
J 0 (ix s ) s 2 J 0 (ia s )
where λ1, λ2,… are the positive roots of J0(λ) = 0 Triangular wave function
33.158.
1 ⎛ as⎞ tanh ⎝ 2⎠ as 2 Fig. 33-1
Square wave function
33.159.
1 ⎛ as⎞ tanh s ⎝ 2⎠ Fig. 33-2
Rectified sine wave function
33.160.
πa ⎛ as⎞ coth ⎝ 2⎠ a2s2 + π 2 Fig. 33-3
Half-rectified sine wave function 33.161.
πa (a s + π 2 )(1 − e − as ) 2 2
Fig. 33-4
Sawtooth wave function 33.162.
2
1 e − as − as 2 s(1 − e − as ) Fig. 33-5
LAPLACE TRANSFORMS
192
f (s)
F(t) Heaviside’s unit function (t – a)
33.163.
e − as s
See also entry 33.138. Fig. 33-6
Pulse function
33.164.
e − as (1 − e − s ) s Fig. 33-7
Step function
33.165.
1 s(1 − e − as ) See also entry 33.102. Fig. 33-8
F(t) = n2, n t < n + 1, n = 0, 1, 2, …
33.166.
e − s + e −2 s s(1 − e − s )2
Fig. 33-9
F(t) = rn, n t < n + 1, n = 0, 1, 2, …
33.167.
1 − e− s s(1 − re − s ) See also entry 33.104. Fig. 33-10
⎧sin (π t /a) 0 t a F (t ) = ⎨ t>a ⎩0 33.168.
π a(1 + e − as ) a2s2 + π 2
Fig. 33-11
34
FOURIER TRANSFORMS
Fourier’s Integral Theorem ∞
f ( x ) = ∫ {A(α ) cos α x + B(α )sin α x}dα
34.1.
0
where
34.2.
1 ∞ ⎧ ⎪A(α ) = π ∫−∞ f ( x ) cos α x dx ⎨ 1 ∞ ⎪ B(α ) = ∫ f ( x ) sin α x dx π −∞ ⎩
Sufficient conditions under which this theorem holds are:
(i) f (x) and f ′(x) are piecewise continuous in every finite interval –L < x < L; ∞ (ii) ∫ | f ( x ) | dx converges; −∞ (iii) f (x) is replaced by 12 { f ( x + 0) + f ( x − 0)} if x is a point of discontinuity.
Equivalent Forms of Fourier’s Integral Theorem 34.3.
f (x) =
1 2π
∫
f (x) =
1 2π
∫
=
1 2π
∫ ∫
2 π
∞
∞
0
0
34.4.
34.5.
f (x) =
∫
∞
α =−∞ ∞
∫
∞ u =−∞
f (u) cos α ( x − u) du dα ∞
e iα x dα ∫ f (u)e − iα u du
−∞
−∞
∞
∞
−∞
−∞
f (u)e iα ( x −u ) du dα
sin α x dα ∫ f (u) sin α u du
where f (x) is an odd function [ f (−x) = −f(x)]. 34.6.
f (x) =
2 π
∫
∞ 0
∞
cos α x dα ∫ f (u) cos α u du 0
where f (x) is an even function [ f (−x) = f (x)].
193
FOURIER TRANSFORMS
194
Fourier Transforms The Fourier transform of f (x) is defined as 34.7.
{ f ( x )} = F (α ) =
∫
∞ −∞
f ( x )e − iα x dx
Then from 34.7 the inverse Fourier transform of F(a ) is 34.8.
−1{F (α )} = f ( x ) =
1 2π
∫
∞ −∞
F (α )e iα x dα
We call f (x) and F(a) Fourier transform pairs.
Convolution Theorem for Fourier Transforms If F(a) = {f (x)} and G(a ) = {g(x)}, then 34.9.
1 2π
∫
∞ −∞
F (α )G (α )eiα x dα =
∫
∞ −∞
f (u)g( x − u) du = f * g
where f *g is called the convolution of f and g. Thus,
34.10. { f *g} = { f} {g}
Parseval’s Identity If F(a) = { f (x)}, then
34.11.
∫
∞
| f ( x ) |2 dx =
−∞
1 2π
∫
∞
| F (α ) |2 dα
−∞
More generally if F(a) = { f (x)} and G(a) = {g(x)}, then
34.12.
∫
∞ −∞
f ( x )g( x ) dx =
1 2π
∫
∞ −∞
F (α )G (α ) dα
where the bar denotes complex conjugate.
Fourier Sine Transforms The Fourier sine transform of f (x) is defined as 34.13.
FS (α ) = S { f ( x )} =
∫
∞ 0
f ( x ) sin α x dx
Then from 34.13 the inverse Fourier sine transform of FS(a ) is
34.14.
f ( x ) = −S1{FS (α )} =
2 π
∫
∞ 0
FS (α ) sin α x dα
FOURIER TRANSFORMS
195
Fourier Cosine Transforms The Fourier cosine transform of f (x) is defined as 34.15.
FC (α ) = C { f ( x )} =
∫
∞ 0
f ( x ) cos α x dx
Then from 34.15 the inverse Fourier cosine transform of FC(a) is
34.16.
f ( x ) = −C1{FC (α )} =
2 π
∫
∞ 0
FC (α ) cos α x dα
Special Fourier Transform Pairs f (x)
F(a )
1 |x|b
2 sin bα α
34.18.
1 x + b2
π e − bα b
34.19.
x x 2 + b2
−iπ e − bα
34.20.
f (n)(x)
inanF(a)
34.21.
x nf (x)
34.22.
f (bx)eitx
34.17.
{
2
in
d nF dα n
1 ⎛ α − t⎞ F b ⎝ b ⎠
FOURIER TRANSFORMS
196
Special Fourier Sine Transforms f (x)
FC(a )
1 0<xb
1 − cos bα α
34.24.
x –1
π 2
34.25.
x x + b2
π − bα e 2
34.26.
e–bx
α α 2 + b2
34.27.
xn – 1e–bx
Γ(n)sin(n tan −1 α / b) (α 2 + b 2 )n / 2
34.28.
xe − bx
34.29.
x –1/2
π 2α
34.30.
x –n
πα n−1 csc (nπ / 2) 2Γ (n)
34.31.
sin bx x
34.32.
sin bx x2
34.33.
cos bx x
α b
34.34.
tan −1 ( x / b)
π − bα e 2α
34.35.
csc bx
π πα tanh 2b 2b
34.36.
1 e −1
π ⎛ πα ⎞ 1 − coth 4 ⎝ 2 ⎠ 2α
34.23.
{
2
2x
2
π α e −α 4 b 3/ 2
2
/4b
0
1 ⎛ α + b⎞ ln 2 ⎝ α − b⎠
{
πα / 2 α < b πb / 2 α > b
FOURIER TRANSFORMS
197
Special Fourier Cosine Transforms f (x)
FC (a )
1 0<xb
sin bα α
34.38.
1 x + b2
π e − bα 2b
34.39.
e − bx
b α 2 + b2
34.40.
x n−1e − bx
Γ(n) cos(n tan −1 α /b) (α 2 + b 2 )n / 2
34.41.
e − bx
34.42.
x −1/ 2
π 2α
34.43.
x −n
πα n −1 sec (nπ /2) , 0 < n <1 2Γ (n)
34.44.
⎛ x 2 + b2 ⎞ ln ⎜ 2 ⎝ x + c 2 ⎟⎠
e − cα − e − bα πα
34.45.
sin bx x
⎧π /2 α < b ⎪ ⎨π /4 α = b ⎪⎩ 0 α > b
34.46.
sin bx 2
π ⎛ α2 α2⎞ cos − sin ⎟ ⎜ 8b ⎝ 4b 4b⎠
34.47.
cos bx 2
π 8b
34.48.
sech bx
34.37.
34.49.
34.50.
{
2
1 π −α e 2 b
2
cosh ( π x/2)
2
/4b
α2 α2⎞ ⎛ ⎜⎝ cos 4 b + sin 4 b ⎟⎠ π πα sech 2b 2b
cosh ( π x )
π cosh ( πα /2) 2 cosh ( πα )
e− b x x
π {cos(2b α ) − sin(2b α )} 2α
Section IX: Elliptic and Miscellaneous Special Functions
35
ELLIPTIC FUNCTIONS
Incomplete Elliptic Integral of the First Kind 35.1.
u = F (k , φ ) =
∫
dθ
φ
1 − k sin θ
0
2
2
=
∫
d
x
(1 − )(1 − k 2 2 )
0
2
where f = am u is called the amplitude of u and x = sin f, and where here and below 0 < k < 1.
Complete Elliptic Integral of the First Kind 35.2.
K = F (k , π / 2) =
∫
dθ
π /2 0
1 − k sin θ 2
2
=
∫
d
1
(1 − )(1 − k 2 2 )
0
2
2 2 2 ⎫⎪ π ⎧⎪ ⎛ 1⎞ 2 ⎛ 1 i 3 ⎞ 4 ⎛ 1 i 3 i 5 ⎞ 6 k + ⎬ = ⎨1 + ⎜ ⎟ k + ⎜ k +⎜ ⎟ ⎟ 2 ⎪ ⎝ 2⎠ ⎝ 2 i 4 i 6⎠ ⎝ 2 i 4⎠ ⎪⎭ ⎩
Incomplete Elliptic Integral of the Second Kind 35.3.
E (k , φ ) =
∫
φ
1 − k 2 sin 2 θ dθ =
0
∫
x
1 − k 2 2
0
1− 2
d
Complete Elliptic Integral of the Second Kind 35.4.
E = E (k , π / 2) =
∫
π /2 0
1 − k 2 sin 2 θ dθ =
∫
1
1 − k 2 2
0
1− 2
d
2 2 2 ⎫⎪ π ⎧⎪ ⎛ 1⎞ 2 ⎛ 1 i 3 ⎞ k 4 ⎛ 1 i 3 i 5 ⎞ k 6 = ⎨1 − ⎜ ⎟ k − ⎜ − − ⎬ 2 ⎪ ⎝ 2⎠ ⎝ 2 i 4 ⎟⎠ 3 ⎜⎝ 2 i 4 i 6⎟⎠ 5 ⎪⎭ ⎩
Incomplete Elliptic Integral of the Third Kind 35.5.
Π(k , n, φ ) =
∫
φ
dθ
0
(1 + n sin θ ) 1 − k sin θ 2
2
2
=
∫
x
d
0
(1 + n ) (1 − 2 )(1 − k 2 2 ) 2
198 Copyright © 2009, 1999, 1968 by The McGraw-Hill Companies, Inc. Click here for terms of use.
ELLIPTIC FUNCTIONS
199
Complete Elliptic Integral of the Third Kind 35.6.
Π(k , n, π / 2) =
∫
dθ
π /2 0
(1 + n sin θ ) 1 − k sin θ 2
2
2
=
∫
1
d
0
(1 + n ) (1 − 2 )(1 − k 2 2 ) 2
Landen’s Transformation 35.7.
tan φ =
sin 2φ1 k + cos 2φ1
k sin φ = sin (2φ1 − φ )
or
This yields 35.8.
F (k , φ ) =
∫
φ 0
2 dθ = 2 2 + k 1 1 − k sin θ
∫
dθ1 1 − k12 sin 2 θ1
φ1 0
where k1 = 2 k /(1 + k ). By successive applications, sequences k1 , k2 , k3 , … and φ1 , φ2 , φ3 , … are obtained such that k < k1 < k2 < k3 < < 1 where lim kn = 1. It follows that n→∞
35.9.
F ( k , Φ) =
k1 k2 k3 … Φ dθ ∫0 1 − sin 2 θ = k
k1 k2 k3 … ⎛ π Φ⎞ ln tan ⎜ + ⎟ k ⎝ 4 2⎠
where 35.10.
k1 =
2 k , 1+ k
k2 =
2 k1 , … and 1 + k1
Φ lim φn n→∞
The result is used in the approximate evaluation of F(k, f).
Jacobi’s Elliptic Functions From 35.1 we define the following elliptic functions: 35.11.
x = sin (am u) = sn u
35.12.
1 − x 2 = cos (am u) = cn u
35.13.
1 − k 2 x 2 = 1 − k 2 sn 2 u = dn u
We can also define the inverse functions sn −1 x , cn −1 x , dn −1 x and the following: 35.14.
ns u =
1 sn u
35.17.
sc u =
sn u cn u
35.20.
cs u =
cn u sn u
35.15.
nc u =
1 cn u
35.18.
sd u =
sn u dn u
35.21.
dc u =
dn u cn u
35.16.
nd u =
1 dn u
35.19.
cd u =
cn u dn u
35.22.
ds u =
dn u dn u
Addition Formulas 35.23.
sn (u + ) =
sn u cn dn + cn u sn dn u 1 − k 2 sn 2 u sn 2
ELLIPTIC FUNCTIONS
200
35.24.
cn (u + ) =
cn u cn − sn u sn dn u dn 1 − k 2sn 2 u sn 2
35.25.
dn (u + ) =
dn u dn − k 2 sn u sn cn u cn 1 − k 2sn 2 u sn 2
Derivatives 35.26.
d sn u = cn u dn u du
35.28.
d dn u = − k 2sn u cn u du
35.27.
d cn u = −sn u dn u du
35.29.
d sc u = dc u nc u du
Series Expansions 35.30.
sn u = u − (1 + k 2 )
35.31.
cn u = 1 −
35.32.
dn u = 1 − k 2
u3 u5 u7 + (1 + 14 k 2 + k 4 ) − (1 + 135k 2 + 135k 4 + k 6 ) + 3! 5! 7!
u2 u4 u6 + (1 + 4 k 2 ) − (1 + 44 k 2 + 16 k 4 ) + 2! 4! 6! u2 u4 u6 + k 2 (4 + k 2 ) − k 2 (16 + 44 k 2 + k 4 ) + 2! 4! 6!
Catalan’s Constant 35.33.
1 1 1 1 π /2 K dk = ∫ ∫ ∫ 2 0 2 k =0 θ =0
1 1 1 dθ dk = 2 − 2 + 2 − = .915965594 … 2 2 1 3 5 1 − k sin θ
Periods of Elliptic Functions Let 35.34.
K=
∫
π /2 0
dθ , 1 − k 2 sin 2 θ
K′ =
Then 35.35.
sn u has periods 4K and 2iK ′
35.36.
cn u has periods 4K and 2K + 2iK ′
35.37.
dn u has periods 2K and 4iK ′
∫
π /2 0
dθ 1 − k ′ 2 sin 2 θ
where k ′ = 1 − k 2
ELLIPTIC FUNCTIONS
201
Identities Involving Elliptic Functions 35.38.
sn 2 u + cn 2u = 1
35.40.
dn 2 u − k 2 cn 2 u = k ′ 2
35.42.
cn 2u =
35.44.
where k ′ = 1 − k 2
dn 2u + cn 2u 1 + dn 2u
1 − cn 2u sn u dn u = 1 + cn 2 u cn u
35.39.
dn 2 u + k 2 sn 2 u = 1
35.41.
sn 2u =
1 − cn 2u 1 + dn 2u
35.43.
dn 2u =
1 − k 2 + dn 2u + k 2cn u 1 + dn 2u
35.45.
1 − dn 2u k sn u cn u = 1 + dn 2u dn u
Special Values 35.46.
sn 0 = 0
35.47. cn 0 = 1
35.48. dn 0 = 1
Integrals 1
35.51.
∫ sn u du = k ln (dn u − k cn u)
35.52.
∫ cn u du = k cos
35.53.
∫ dn u du = sin
35.54.
∫ sc u du =
35.55.
∫ cs u du = ln (ns u − ds u)
35.56.
∫ cd u du = k ln (nd u + k sd u)
35.57.
∫ dc u du = ln (nc u + sc u)
35.58.
∫ sd u du = k
35.59.
∫ ds u du = ln (ns u − cs u)
35.60.
∫ ns u du = ln (ds u − cs u)
35.61.
∫ nc u du =
⎛ 1 sc u ⎞ ln ⎜ dc u + 2 ⎝ 1− k 1 − k 2 ⎟⎠
35.62.
∫ nd u du =
1 cos −1 (cd u) 1 − k2
1
−1
−1
(dn u)
(sn u)
(
)
1 ln dc u + 1 − k 2 nc u 1 − k2
1
−1 sin −1 (k cd u) 1 − k2
35.49. sc 0 = 0
35.50. am 0 = 0
ELLIPTIC FUNCTIONS
202
Legendre’s Relation 35.63.
EK ′ + E ′K − KK ′ = π /2
where 35.64.
E=
35.65.
E′ =
∫
π /2 0
∫
π /2 0
1 − k 2 sin 2 θ dθ 1 − k ′ 2 sin 2 θ dθ
K=
∫
K′ =
∫
π /2 0
π /2 0
dθ 1 − k 2 sin 2 θ dθ 1 − k ′ 2 sin 2 θ
MISCELLANEOUS and RIEMANN ZETA FUNCTIONS
36
2 π
Error Function erf ( x) = 36.1.
∫
x
0
e− u du 2
x3 x5 x7 2 ⎛ ⎞ ⎜⎝ x − 3 i 1! + 5 i 2! − 7 i 3! + ⎟⎠ π e− x ⎛ 1 1i 3 1i 3 i 5 ⎞ erf ( x ) ~ 1 − ⎜⎝1 − 2 x 2 + (2 x 2 )2 − (2 x 2 )3 + ⎟⎠ πx erf ( x ) =
2
36.2. 36.3.
erf (− x ) = − erf ( x ),
erf (0) = 0,
erf (∞) = 1
Complementary Error Function erfc ( x) = 1 − erf ( x) = 36.4.
erfc ( x ) = 1 −
36.5.
erfc ( x ) ~
36.6.
erfc (0) = 1,
2 π
∫
∞ x
e− u du 2
x3 x5 x7 ⎛ ⎞ ⎜⎝ x − 3 i 1! + 5 i 2! − 7 i 3! + ⎟⎠
2 π
e− x ⎛ 1 1i 3 1i 3i 5 ⎞ ⎜1 − 2 + (2 x 2 )2 − (2 x 2 )3 + ⎟⎠ π x ⎝ 2x 2
erfc (∞) = 0 ∞ −u
Exponential Integral Ei ( x) = ∫ e du x u 36.7. 36.8. 36.9. 36.10.
Ei ( x ) = −γ − ln x + ∫
x 0
1 − e−u du u
x2 x3 ⎛ x ⎞ Ei ( x ) = −γ − ln x + ⎜ − + − ⎟ ⎝ 1 i 1! 2 i 2! 3 i 3! ⎠ e − x ⎛ 1! 2! 3! ⎞ Ei ( x ) ~ 1 − + 2 − 3 + ⎟ x ⎜⎝ x x x ⎠ Ei (∞) = 0 x
Sine Integral Si ( x) = ∫0
sin u du u
36.11.
Si ( x ) =
x x3 x5 x7 − + − + 1 i 1! 3 i 3! 5 i 5! 7 i 7!
36.12.
Si ( x ) ~
π sin x ⎛ 1 3! 5! 2! 4 ! ⎞ ⎞ cos x ⎛ + − ⎟ − − + − ⎟ − 1− 2 x ⎜⎝ x x 3 x 5 x ⎜⎝ x 2 x 4 ⎠ ⎠
36.13.
Si (− x ) = −Si ( x ),
Si (0) = 0,
Si (∞) = π / 2
203
MISCELLANEOUS AND RIEMANN ZETA FUNCTIONS
204 ∞
Cosine Integral Ci( x) = ∫x 36.14.
Ci ( x ) = −γ − ln x + ∫
36.15.
Ci ( x ) = −γ − ln x +
36.16.
Ci ( x ) ~
36.17.
Ci (∞) = 0
x 0
cos u du u
1 − cos u du u
x2 x4 x6 x8 − + − + 2 i 2! 4 i 4 ! 6 i 6! 8 i 8!
cos x ⎛ 1 3! 5! 2! 4 ! ⎞ sin x ⎛ ⎞ − + − ⎟ − + − ⎟ 1− x ⎜⎝ x x 3 x 5 x ⎜⎝ x 2 x 4 ⎠ ⎠
2
∫
Fresnel Sine Integral S( x) = π 36.18. 36.19. 36.20.
x
0
2 ⎛ x3 x7 x 11 x 15 ⎞ − + − + ⎟ ⎜ π ⎝ 3 i 1! 7 i 3! 11 i 5! 15 i 7! ⎠ i i i i 1i 3i 5 1 1 ⎧ 1 1 3 1 3 5 7 ⎞ ⎞⎫ 2 ⎛ 2 ⎛ 1 S(x) ~ − ⎨(cos x ) ⎜⎝ x − 22 x 5 + 24 x 9 − ⎟⎠ + (sin x ) ⎜⎝ 2 x 3 − 23 x 7 + ⎟⎠ ⎬ 2 2π ⎩ ⎭ S(x) =
S (− x ) = − S ( x ),
S(0) = 0,
S(∞) = 2
Fresnel Cosine Integral C( x) = π 36.21. 36.22. 36.23.
sin u2 du
∫
x
0
1 2
cos u2 du
x5 x9 x 13 2 ⎛x ⎞ − + − + ⎟ ⎜ π ⎝ 1! 5 i 2! 9 i 4 ! 13 i 6! ⎠ 1i 3i 5 1 1 ⎧ 1i 3 1i 3i 5i 7 ⎞ ⎞⎫ 2 ⎛1 2 ⎛ 1 C(x) ~ + ⎨(sin x ) ⎜⎝ x − 22 x 5 + 24 x 9 − ⎟⎠ − (cos x ) ⎜⎝ 2 x 3 − 23 x 7 + ⎟⎠ ⎬ 2 2π ⎩ ⎭ C(x) =
C (− x ) = −C ( x ),
C(0) = 0,
C(∞) = 1
1
1 2
1
Riemann Zeta Function ζ ( x) = 1x + 2 x + 3 x + ∞ u x −1 1 du, Γ ( x ) ∫0 eu−1
36.24.
ζ (x) =
36.25.
ζ (1 − x ) = 21− x π − x Γ ( x ) cos(π x/ 2)ζ ( x ) (extension to other values)
36.26.
ζ (2 k ) =
x >1
22 k −1 π 2 k Bk k = 1, 2, 3,… (2k )!
Section X: Inequalities and Infinite Products
37
INEQUALITIES
Triangle Inequality 37.1.
|a1 | − | a2 | | a1 + a2 | | a1 | + | a2 |
37.2.
| a1 + a2 + + an | | a1 | + | a2 | + + | an |
Cauchy-Schwarz Inequality 37.3.
(a1b1 + a2 b2 + + an bn )2 (a12 + a22 + + an2 )(b12 + b22 + + bn2 )
The equality holds if and only if a1 /b1 = a2 /b2 = = an /bn .
Inequalities Involving Arithmetic, Geometric, and Harmonic Means If A, G, and H are the arithmetic, geometric, and harmonic means of the positive numbers a1, a2, ..., an, then 37.4.
H G A
where a1 + a2 + + an n
37.5.
A=
37.6.
G = n a1 a2 … an
37.7.
1 1⎛1 1 1⎞ = ⎜ + ++ ⎟ H n ⎝ a1 a2 an ⎠
The equality holds if and only if a1 = a2 = = an .
Holder’s Inequality 37.8.
| a1b1 + a2b2 + + an bn | (| a1 | p+ | a2 | p + + | an | p)1/ p (| b1 |q + | b2 |q + + | bn |q)1/q
where 37.9.
1 1 + =1 p q
p > 1, q > 1
The equality holds if and only if | a1 | p−1/| b1 | = | a2 | p−1/| b2 | = = | an | p−1/| bn | . For p = q = 2 it reduces to 37.3.
205 Copyright © 2009, 1999, 1968 by The McGraw-Hill Companies, Inc. Click here for terms of use.
INEQUALITIES
206
Chebyshev’s Inequality If a1 a2 an and b1 b2 bn , then 37.10.
⎛ a1 + a2 + + an ⎞ ⎛ b1 + b2 + + bn ⎞ a1 b1 + a2 b2 + + an bn ⎟ ⎟⎜ ⎜ n n n ⎠ ⎠⎝ ⎝
or 37.11.
(a1 + a2 + + an )(b1 + b2 + + bn ) n(a1 b1 + a2 b2 + + an bn )
Minkowski’s Inequality If a1, a2,…an, b1, b2,… bn are all positive and p > 1, then 37.12.
{(a1 + b1) p + (a2 + b2 ) p + + (an + bn ) p}1/ p (a1p + a2p + + anp )1/ p + (b1p + b2p + + bnp )1/ p
The equality holds if and only if a1 /b1 = a2 /b2 = = an /bn .
Cauchy-Schwarz Inequality for Integrals 2
{
}{
}
b ⎡ b f ( x )g( x ) dx⎤ b [ f ( x )]2 dx ∫a ∫a [g( x)]2 dx ⎢⎣∫a ⎥⎦ The equality holds if and only if f (x)/g(x) is a constant.
37.13.
Holder’s Inequality for Integrals 37.14.
∫
b
a
| f ( x )g( x )|dx
{∫ | f (x)| dx} {∫ |g(x)| dx} b
1/ p
1/q
b
p
q
a
a
where 1/p + 1/q = 1, p > 1, q >1. If p = q = 2, this reduces to 37.13. The equality holds if and only if | f ( x )| p−1/| g( x )| is a constant.
Minkowski’s Inequality for Integrals If p > 1, 37.15.
{∫ | f (x) + g(x)| dx} {∫ | f (x)| dx} + {∫ |g(x)| dx} b
a
1/ p
p
b
a
1/ p
p
b
a
The equality holds if and only if f (x)/g(x) is a constant.
1/ p
p
38
INFINITE PRODUCTS
38.1.
⎛ x2 ⎞ ⎛ x2 ⎞ ⎛ x2 ⎞ sin x = x ⎜1 − 2 ⎟ ⎜1 − 2 ⎟ ⎜1 − 2 ⎟ ⎝ x ⎠ ⎝ 4π ⎠ ⎝ 9π ⎠
38.2.
⎛ 4x2 ⎞ ⎛ 4x2 ⎞ ⎛ 4x2 ⎞ cos x = ⎜1 − 2 ⎟ ⎜1 − 2 ⎟ ⎜1 − π ⎠ ⎝ 9π ⎠ ⎝ 25π 2 ⎟⎠ ⎝
38.3.
⎛ x2 ⎞ ⎛ x2 ⎞ ⎛ x2 ⎞ sinh x = x ⎜1 + 2 ⎟ ⎜1 + 2 ⎟ ⎜1 + 2 ⎟ ⎝ π ⎠ ⎝ 4π ⎠ ⎝ 9π ⎠
38.4.
⎛ 4x2 ⎞ ⎛ 4x2 ⎞ ⎛ 4x2 ⎞ cosh x = ⎜1 + 2 ⎟ ⎜1 + 2 ⎟ ⎜1 + π ⎠ ⎝ 9π ⎠ ⎝ 25π 2 ⎟⎠ ⎝
38.5.
1 ⎫ ⎧⎛ x ⎫ ⎧⎛ x ⎫ ⎧⎛ x = xeγ x ⎨ 1 + ⎞ e − x ⎬ ⎨ 1 + ⎞ e − x/2 ⎬ ⎨ 1 + ⎞ e − x/3 ⎬ 1 2 Γ( x ) 3 ⎝ ⎠ ⎝ ⎠ ⎝ ⎠ ⎭⎩ ⎭⎩ ⎭ ⎩
See also 25.11. ⎛ x2 ⎞ ⎛ x2 ⎞ ⎛ x2 ⎞ J 0 ( x ) = ⎜1 − 2 ⎟ ⎜1 − 2 ⎟ ⎜1 − 2 ⎟ ⎝ 1 ⎠ ⎝ 2 ⎠ ⎝ 3 ⎠ where 1, 2, 3,… are the positive roots of J0(x) = 0.
38.6.
⎛ x2 ⎞ ⎛ x2 ⎞ ⎛ x2 ⎞ J1( x ) = x ⎜1 − 2 ⎟ ⎜1 − 2 ⎟ ⎜1 − 2 ⎟ ⎝ 1 ⎠ ⎝ 2 ⎠ ⎝ 3 ⎠ where 1, 2, 3,… are the positive roots of J1(x) = 0.
38.7.
38.8.
sin x x x x x = cos cos cos cos 2 4 8 16 x
38.9.
π 2 2 4 4 6 6 = i i i i i i 2 1 3 3 5 5 7
This is called Wallis’ product.
207
Section XI: Probability and Statistics
39
DESCRIPTIVE STATISTICS
The numerical data x1, x2,… will either come from a random sample of a larger population or from the larger population itself. We distinguish these two cases using different notation as follows: n = number of items in a sample, N = number of items in the population, x = (read: x-bar) = sample mean, s2 = sample variance, s = sample standard deviation,
m (read: mu) = population mean, s 2 = population variance, s = population standard deviation
Note that Greek letters are used with the population and are called parameters, whereas Latin letters are used with the samples and are called statistics. First we give formulas for the data coming from a sample. This is followed by formulas for the population. Grouped Data Frequently, the sample data are collected into groups (grouped data). A group refers to a set of numbers all with the same value xi, or a set (class) of numbers in a given interval with class value xi. In such a case, we assume there are k groups with fi denoting the number of elements in the group with value or class value xi. Thus, the total number of data items is 39.1.
n = ∑ fi
As usual, Σ will denote a summation over all the values of the index, unless otherwise specified. Accordingly, some of the formulas will be designated as (a) or as (b), where (a) indicates ungrouped data and (b) indicates grouped data.
Measures of Central Tendency Mean (Arithmetic Mean) The arithmetic mean or simply mean of a sample x1, x2,…, xn, frequently called the “average value,” is the sum of the values divided by the number of values. That is: 39.2(a). Sample mean: 39.2(b). Sample mean:
x= x=
x1 + x 2 + + x n Σ xi = n n f1 x1 + f2 x 2 + + fk x k Σ fi xi = Σ fi f1 + f2 + + fk
Median Suppose that the data x1, x2,…, xn are now sorted in increasing order. The median of the data, denoted by M or Median
is defined to be the “middle value.” That is:
208 Copyright © 2009, 1999, 1968 by The McGraw-Hill Companies, Inc. Click here for terms of use.
DESCRIPTIVE STATISTICS
209
when n is odd and n = 2k + 1, ⎧x k +1 ⎪ 39.3(a). Median = ⎨ x + x k +1 ⎪ k when n is even and n = 2k. ⎩ 2 The median of grouped data is obtained by first finding the cumulative frequency function Fs. Specifically, we define Fs = f1 + f2 + + fs
that is, Fs is the sum of the frequencies up to fs. Then: 39.3(b.1).
⎧x j +1 ⎪ Median = ⎨ x + x j +1 ⎪ j ⎩ 2
when n = 2k + 1 (odd) and Fj < k + 1 ≤ Fj +1 when n = 2k (even), and Fj = k.
Finding the median of data arranged in classes is more complicated. First one finds the median class m, the class with the median value, and then one linearly interpolates in the class using the formula 39.3(b.2).
Median = Lm + c
(n/2) − Fm −1 fm
where Lm denotes the lower class boundary of the median class and c denotes its class width (length of the class interval). Mode The mode is the value or values which occur most often. Namely: 39.4. Mode xm = numerical value that occurs the most number of times The mode is not defined if every xm occurs the same number of times, and when the mode is defined it may not be unique. Weighted and grand means Suppose that each xi is assigned a weight wi ≥ 0. Then: 39.5. Weighted Mean x w =
w1x1 + w2 x 2 + + wk x k Σ wi xi = Σ wi w1 + w2 + + wk
Note that 39.2(b.1) is a special case of 39.4 where the weight wi of xi is its frequency. Suppose that there are k sample sets and that each sample set has ni elements and a mean x. Then the grand mean, denoted by xi is the “mean of the means” where each mean is weighted by the number of elements in its sample. Specifically: 39.6. Grand Mean x =
n1 x1 + n2 x 2 + + nk x k Σ ni xi = Σ ni n1 + n2 + + nk
Geometric and Harmonic Means The geometric mean (G.M.) and harmonic mean (H.M.) are defined as follows: 39.7(a). G.M. = n x1 x 2 x n f f f 39.7(b). G.M. = n x1 x 2 x k 1
2
k
DESCRIPTIVE STATISTICS
210
39.8(a).
H.M. =
39.8(b).
H.M. =
n n = 1/x1 + 1/x 2 + + 1/x n Σ (1/xi ) n n = f1 /x1 + f2 /x 2 + + fk /x k Σ ( fk /xi )
Relation Between Arithmetic, Geometric, and Harmonic Means 39.9. H.M. ≤ G.M. ≤ x The equality sign holds only when all the sample values are equal. Midrange The midrange is the average of the smallest value x1 and the largest value xn. That is: 39.10. midrange: mid =
x1 + x n 2
Population Mean The formula for the population mean m follows: 39.11(a). Population mean:
x1 + x 2 + + x N Σ xi = N N
μ=
39.11(b). Population mean: μ =
f1 x1 + f2 x 2 + + fk x k Σ fi xi = Σ fi f1 + f2 + + fk
(Recall that N denotes the number of elements in a population.) Observe that the formula for the population mean m is the same as the formula for the sample mean x. On the other hand, the formula for the population standard deviation s is not the same as the formula for the sample standard deviation s. (This is the main reason we give separate formulas for m and x. )
Measures of Dispersion Sample Variance and Standard Deviation Here the sample set has n elements with mean x. 39.12(a). Sample variance:
s2 =
Σ( xi − x )2 Σ xi2 − (Σ xi )2 /n = n −1 n −1
39.12(b). Sample variance:
s2 =
Σfi ( xi − x )2 Σfi xi2 − (Σ fi xi )2 / Σ fi = ( Σ fi ) − 1 ( Σ fi ) − 1 s = Variance = s 2
39.13. Sample standard deviation: EXAMPLE 39.1:
Consider the following frequency distribution: xi
1
2
3
4
5
6
fi
8
14
7
12
3
1
Then n = Σ fi = 45 and Σ fi xi = 126. Hence, by 39.2(b), Mean x =
Σ xi fi 126 = = 2.8 Σ fi 45
DESCRIPTIVE STATISTICS
211
Also, n – 1 = 44 and Σ fi xi2 = 430. Hence, by 39.12(b) and 39.13, s2 =
430 − (126)2 /45 ≈ 1.75 and s = 1.32 44
We find the median M, first finding the cumulative frequencies: F1 = 8,
F2 = 22,
F3 = 29,
F4 = 41,
F5 = 44,
F6 = 45 = n
Here n is odd, and (n + 1)/2 = 23. Hence, Median M = 23rd value = 3 The value 2 occurs most often, hence Mode = 2 M.D. and R.M.S. Here M.D. stands for mean deviation and R.M.S. stands for root mean square. As previously, x is the mean of the data and, for grouped data, n = Σ fi. 1 x −x n i
39.14(a).
M.D. =
39.15(a).
R.M.S. =
1 (Σ xi2 ) n
1 f x −x n i i
39.14(b).
M.D. =
39.15(b).
R.M.S. =
1 (Σ fi xi2 ) n
Measures of Position (Quartiles and Percentiles) Now we assume that the data x1, x2,…, xn are arranged in increasing order. 39.16. Sample range: xn – x1. There are three quartiles: the first or lower quartile, denoted by Q1 or QL; the second quartile or median, denoted by Q2 or M; and the third or upper quartile, denoted by Q3 or QU. These quartiles (which essentially divide the data into “quarters”) are defined as follows, where “half” means n/2 when n is even and (n-1)/2 when n is odd: 39.17. QL(= Q1) = median of the first half of the values. M (= Q2 ) = median of the values. QU (= Q3) = median of the second half of the values. 39.18. Five-number summary: [L, QL, M, QU, H] where L = x1 (lowest value) and H = xn (highest value). 39.19. Innerquartile range: QU – QL 39.20. Semi-innerquartile range:
Q=
QU − QL 2
The kth percentile, denoted by Pk, is the number for which k percent of the values are at most Pk and (100–k) percent of the values are greater than Pk. Specifically: 39.21. Pk = largest xs such that Fs ≤ k/100. Thus, QL = 25th percentile, M = 50th percentile, QU = 75th percentile.
DESCRIPTIVE STATISTICS
212
Higher-Order Statistics 39.22. The rth moment: (a) mr =
1 Σ xir , n
(b) mr =
1 Σ f xr n i i
39.23. The rth moment about the mean x: (a)
μr =
1 Σ ( x i − x )r , n
(b)
μr =
1 Σ ( fi x i − x ) r n
μr =
1 Σ fi x i − x n
39.24. The rth absolute moment about mean x: (a)
μr =
r 1 Σ xi − x , n
(b)
r
39.25. The rth moment in standard z units about z = 0:
αr =
1 r Σz , n i
x −x 1 Σ f z r where zi = i σ n i i
(b)
αr =
39.26. Coefficient of skewness:
γ1 =
μ3 = α3 σ3
39.27. Momental skewness:
μ3 2σ 3
39.28. Coefficient of kurtosis:
α4 =
39.29. Coefficient of excess (kurtosis):
α4 − 3 =
39.30. Quartile coefficient of skewness:
QU − 2 xˆ + QL Q3 − 2Q2 + Q1 = QU − QL Q3 − Q1
(a)
Measures of Skewness and Kurtosis
μ4 σ4 μ4 −3 σ4
Population Variance and Standard Deviation Recall that N denotes the number of values in the population. 39.31. Population variance: σ 2 =
Σ ( xi − x )2 Σ xi2 − (Σ xi )2 /n = N N
39.32. Population standard deviation: σ = Variance = σ 2
Bivariate Data The following formulas apply to a list of pairs of numerical values: ( x1, y1), ( x 2 , y2 ), ( x3, y3),… , ( x n , yn )
where the first values correspond to a variable x and the second to a variable y. The primary objective is to determine whether there is a mathematical relationship, such as a linear relationship, between the data. The scatterplot of the data is simply a picture of the pairs of values as points in a coordinate plane.
DESCRIPTIVE STATISTICS
213
Correlation Coefficient A numerical indicator of a linear relationship between variables x and y is the sample correlation coefficient r of x and y, defined as follows: r=
39.33. Sample correlation coefficient:
Σ ( xi − x )( yi − y ) Σ ( xi − x )2 Σ ( yi − y )2
We assume that the denominator in Formula 39.33 is not zero. An alternative formula for computing r follows: 39.34.
r=
Σ xi yi − (Σ xi )(Σ yi)/n Σ xi2 − (Σ xi )2 /n Σ yi2 − (Σ yi)2 /n
Properties of the correlation coefficient r follow: 39.35. (1) –1 r 1 or, equivalently, r 1. (2) r is positive or negative according as y tends to increase or decrease as x increases. (3) The closer |r| is to 1, the stronger the linear relationship between x and y. The sample covariance of x and y is denoted and defined as follows: sxy =
39.36. Sample covariance:
Σ ( xi − x )( yi − y ) n −1
Using the sample covariance, Formula 39.33 can be written in the compact form: 39.37.
r=
sxy sx s y
where sx and sy are the sample standard deviations of x and y, respectively. EXAMPLE 39.2:
Consider the following data: x y
50 2.5
45 5.0
40 6.2
38 7.4
32 8.3
40 4.7
55 1.8
The scatterplot of the data appears in Fig. 39-1. The correlation coefficient r for the data may be obtained by first constructing the table in Fig. 39-2. Then, by Formula 39.34 with n = 7,
r=
1431.8 − (300)(35.9) / 7 ≈ −0.9562 13, 218 − (300)2 / 7 218.67 + (35.9)2 / 7
Here r is close to –1, and the scatterplot in Fig. 39-1 does indicate a strong negative linear relationship between x and y.
DESCRIPTIVE STATISTICS
214
Fig. 39-1
Fig. 39-2
Regression Line Consider a given set of n data points Pi (xi, yi). Any (nonvertical) line L may be defined by an equation of the form y = a + bx Let yi * denote the y value of the point on L corresponding to xi; that is, let yi* = a + bxi . Now let di = yi − yi* = yi − (a + bxi )
that is, di is the vertical (directed) distance between the point Pi and the line L. The squares error between the line L and the data points is defined by 39.38.
Σ di2 = d12 + d 22 + + d n2
The least-squares line or the line of best fit or the regression line of y on x is, by definition, the line L whose squares error is as small as possible. It can be shown that such a line L exists and is unique. The constants a and b in the equation y = a + bx of the line L of best fit can be obtained from the following two normal equations, where a and b are the unknowns and n is the number of points: 39.39.
⎧⎪ na + (Σ xi ) b = Σ yi ⎨ ⎪⎩(Σ xi )a + (Σ xi2 )b = Σ xi yi
The solution of the above normal equations follows: 39.40.
b=
n Σ xi yi − (Σ xi )(Σ yi ) rs y ; = n Σ xi2 − (Σ xi )2 sx
a=
Σ yi Σ xi −b = y − bx n n
The second equation tells us that the point ( x , y ) lies on L, and the first equation tells us that the point ( x + sx , y + rs y ) also lies on L. EXAMPLE 39.3: Suppose we want the line L of best fit for the data in Example 39.2. Using the table in Fig. 39-2 and n = 7, we obtain the normal equations
7a + 300 b = 35.9 300 a + 13, 218b = 1431.8
Substitution in 39.40 yields b=
7(1431.8) − (300)(35.9) = − 0.22959 7(13, 218) − (300)2
a=
35.9 300 − (−0.2959) = 17.8100 7 7
DESCRIPTIVE STATISTICS
215
Thus, the line L of best fit is y = 17.8100 – 0.2959x The graph of L appears in Fig. 39-3.
Fig. 39-3
Curve Fitting Suppose that n data points Pi (xi, yi) are given, and that the data (using the scatterplot or the correlation coefficient r) do not indicate a linear relationship between the variables x and y, but do indicate that some other standard (well-known) type of curve y = f (x) approximates the data. Then the particular curve C that one uses to approximate that data, called the best-fitting or least-squares curve, is the curve in the collection which minimizes the squares error sum Σ di2 = d12 + d 22 + + d n2 where di = yi – f(xi). Three such types of curve are discussed as follows. Polynomial function of degree m: y = a0 + a1 x + a2 x 2 + + am x m The coefficients a0 , a1, a2 ,… , am of the best-fitting polynomial can be obtained by solving the following system of m + 1 normal equations: na0 + a1Σ xi + a2 Σ xi2 + + am Σ xim = Σ yi
39.41.
a0 Σ xi + a1 Σ xi2 + a2 Σ xi3 + + am Σ xim +1 = Σ xi yi .................................................................................... a0 Σ xim + a1 Σ xim +1 + a2 Σ xim + 2 + + am Σ xi2 m = Σ xim yi
Exponential curve:
y = ab x or log y = log a + (log b) x
The exponential curve is used if the scatterplot of log y verses x indicates a linear relationship. Then log a and log b are obtained from transformed data points. Namely, the best-fit line L for data points P′(xi, log yi) is 39.42.
⎧⎪ na ′ + (Σ xi ) b ′ = Σ (log yi ) ⎨ ⎪⎩(Σ xi ) a ′ + (Σ xi2 ) b ′ = Σ ( xi log yi )
Then a = antilog a′, b = antilog b′. EXAMPLE 39.4:
Consider the following data which indicates exponential growth: x y
1 6
2 18
3 55
4 160
5 485
6 1460
DESCRIPTIVE STATISTICS
216
Thus, we seek the least-squares line L for the following data: x log y
1 0.7782
2 1.2553
3 1.7404
4 2.2041
5 2.6857
6 3.1644
Using the normal equation 39.42 for L, we get a ′ = 0.3028, b ′ = 0.4767
The antiderivatives of a′ and b′ yield, approximately, a = 2.0,
b = 3.0
Hence, y = 2(3 ) is the required exponential curve C. The data points and C are depicted in Fig. 39-4. x
Fig. 39-4
Power function: y = axb or log y = log a + b log x The power curve is used if the scatterplot of log y verses log x indicates a linear relationship. The log a and b are obtained from transformed data points. Namely, the best-fit line L for transformed data points P′(log xi, log yi) is 39.43.
⎧⎪ na ′ + Σ (log xi )b = Σ (log yi ) ⎨ ⎪⎩Σ (log xi )a ′ + Σ (log xi )2 b = Σ (log xi log yi )
Then a = antilog a′.
40
PROBABILITY
Sample Spaces and Events Let S be a sample space which consists of the possible outcomes of an experiment where the events are subsets of S. The sample space S itself is called the certain event, and the null set ∅ is called the impossible event. It would be convenient if all subsets of S could be events. Unfortunately, this may lead to contradictions when a probability function is defined on the events. Thus, the events are defined to be a limited collection C of subsets of S as follows. DEFINITION 40.1: properties:
The class C of events of a sample space S form a σ-field. That is, C has the following three
(i) S ∈ C. (ii) If A1, A2, … belong to C, then their union A1 ∪ A2 ∪ A3 ∪ … belongs to C. (iii) If A ∈ C, then its complement Ac ∈ C.
Although the above definition does not mention intersections, DeMorgan’s law (40.3) tells us that the complement of a union is the intersection of the complements. Thus, the events form a collection that is closed under unions, intersections, and complements of denumerable sequences. If S is finite, then the class of all subsets of S form a σ-field. However, if S is nondenumerable, then only certain subsets of S can be the events. In fact, if B is the collection of all open intervals on the real line R, then the smallest σ-field containing B is the collection of Borel sets in R. If Condition (ii) in Definition 40.1 of a σ-field is replaced by finite unions, then the class of subsets of S is called a field. Thus a σ-field is a field, but not visa versa. First, for completeness, we list basic properties of the set operations of union, intersection, and complement. 40.1.
Sets satisfy the properties in Table 40-1.
TABLE 40-1
Laws of the Algebra of Sets
Idempotent laws:
(1a) A ∪ A = A
Associative laws:
(2a)
(A ∪ B) ∪ C = A ∪ (B ∪ C)
Commutative laws: (3a) A ∪ B = B ∪ A
(1b) A ∩ A = A (2b) (A ∩ B) ∩ C = A ∩ (B ∩ C) (3b) A ∩ B = B ∩ A
Distributive laws: (4a) A ∪ (B ∩ C) = (A ∪ B) ∩ (A ∪ C) (4b) A ∩ (B ∪ C) = (A ∩ B) ∪ (A ∩ B) Identity laws:
(5a) A ∪ ∅ = A (6a) A ∪ U = U
Involution law:
(7)
(5b) A ∩ U = A (6b) A ∩ ∅ = ∅
(AC)C = A
Complement laws: (8a) A ∪ Ac = U (9a) Uc = ∅
(8b) A ∩ Ac = ∅ (9b) ∅c = U
DeMorgan’s laws: (10a) (A ∪ B)c = Ac ∩ Bc
(10b) (A ∩ B)c = Ac ∪ Bc
217
PROBABILITY
218
40.2. The following are equivalent: (i) A ⊆ B, (ii) A ∩ B = A, (iii) A ∩ B = B. Recall that the union and intersection of any collection of sets is defined as follows: ∪j Aj = {x | there exists j such that x ∈ Aj} 40.3.
and
∩j Aj = {x | for every j we have x ∈ Aj}
(Generalized DeMorgan’s Law) (10a)'(∪j Aj)c = ∩j Ajc; (10b)'(∩j Aj)c = ∪j Ajc
Probability Spaces and Probability Functions DEFINITION 40.2: Let P be a real-valued function defined on the class C of events of a sample space S. Then P is called a probability function, and P(A) is called the probability of an event A, when the following axioms hold:
Axiom [P1] For every event A, P(A) ≥ 0. Axiom [P2] For the certain event S, P(S) = 1. Axiom [P3] For any sequence of mutually exclusive (disjoint) events A1, A2, …,
P(A1 ∪ A2 ∪ …) = P(A1) + P(A2) + … The triple (S, C, P), or simply S when C and P are understood, is called a probability space. Axiom [P3] implies an analogous axiom for any finite number of sets. That is: Axiom [P3'] For any finite collection of mutually exclusive events A1, A2, …, An, P(A1 ∪ A2 ∪ … ∪ An) = P(A1) + P(A2) + … + P(An) In particular, for two disjoint events A and B, we have P(A ∪ B ) = P(A) + P(B). The following properties follow directly from the above axioms. 40.4.
(Complement rule) P(Ac) = 1 – P(A). Thus, P(∅) = 0.
40.5.
(Difference Rule) P(A\B) = P(A) – P(A ∩ B).
40.6.
(Addition Rule) P(A ∪ B) = P(A) + P(B) – P(A ∩ B).
40.7. For n ≥ 2, P 40.8.
(∪
n j =1
)
Aj ≤
∑
n j =1
P( Aj )
(Monoticity Rule) If A ⊆ B, then P(A) ≤ P(B).
Limits of Sequences of Events 40.9. (Continuity) Suppose A1, A2, … form a monotonic increasing (decreasing) sequence of events; that is, Aj ⊆ Aj+1 (Aj ⊇ Aj+1). Let A = ∪ j Aj (A = ∩ j Aj). Then lim P(An) exists and lim P(An) = P(A) For any sequence of events A1, A2, …, we define lim inf An =
∪ ∩ +∞
+∞
k=1
j=k
Aj
and
lim sup An =
∩ ∪ +∞
+∞
k=1
j=k
Aj
If lim inf An = lim sup An, then we call this set lim An. Note lim An exists when the sequence is monotonic.
PROBABILITY
40.10.
219
For any sequence Aj of events in a probability space, P(lim inf An) ≤ lim inf P(An) ≤ lim sup P(An) ≤ P(lim sup An) Thus, if lim An exists, then P(lim An) = lim P(An).
40.11.
For any sequence Aj of events in a probability space, P(∪j Aj) ≤
∑
j
P(Aj).
40.12. (Borel-Cantelli Lemma) Suppose Aj is any sequence of events in a probability space. Furthermore, +∞ suppose ∑ n=1 P(An) < +∞. Then P(lim sup An) = 0. 40.13. (Extension Theorem) Let F be a field of subsets of S. Let P be a function on F satisfying Axioms P1, P2, and P3¢. Then there exists a unique probability function P* on the smallest σ-field containing F such that P* is equal to P on F.
Conditional Probability DEFINITION 40.3: defined as follows:
Let E be an event with P(E) > 0. The conditional probability of an event A given E is denoted and
P(A|E) =
40.14.
P( A ∩ E ) P(E )
(Multiplication Theorem for Conditional Probability) P(A ∩ B) = P(A)P(B|A). This theorem can be genealized as follows:
40.15. P(A1 ∩ … ∩ An) = P(A1)P(A2|A1)P(A3|A1 ∩ A2) … P(An|A1 ∩ … ∩ An-1) EXAMPLE 40.1: A lot contains 12 items of which 4 are defective. Three items are drawn at random from the lot one after the other. Find the probabiliy that all three are nondefective.
The probability that the first item is nondefective is 8/12. Assuming the first item is nondefective, the probability that the second item is nondefective is 7/11. Assuming the first and second items are nondefective, the probability that the third item is nondefective is 6/10. Thus, p=
8 7 6 14 ⋅ ⋅ = 12 11 10 55
Stochastic Processes and Probability Tree Diagrams A (finite) stochastic process is a finite sequence of experiments where each experiment has a finite number of outcomes with given probabilities. A convenient way of describing such a process is by means of a probability tree diagram, illustrated below, where the multiplication theorem (40.14) is used to compute the probability of an event which is represented by a given path of the tree. EXAMPLE 40.2: Let X, Y, Z be three coins in a box where X is a fair coin, Y is two-headed, and Z is weighted so the probability of heads is 1/3. A coin is selected at random and is tossed. (a) Find P(H), the probability that heads appears. (b) Find P(X|H), the probability that the fair coin X was picked if heads appears.
PROBABILITY
220
The probability tree diagram corresponding to the two-step stochastic process appears in Fig. 40-1a. (a) Heads appears on three of the paths (from left to right); hence, P(H) =
1 1 1 1 1 11 ⋅ + ⋅1 + ⋅ = 3 2 3 3 3 18
(b) X and heads H appear only along the top path; hence P(X ∩ H) =
1 1 1 P( X ∩ H ) 1/ 6 3 ⋅ = and so P(X|H) = = = 3 2 6 P( H ) 11 / 18 11
1/2
3%
H A
X 1/2
1/3 1/3 o
D
50%
T
30%
1 H
Y
N
o
4%
D
5%
D
B N
1/3
1/3
H
2/3
T
20%
Z
C N (b)
(a)
Fig. 40-1
Law of Total Probability and Bayes’ Theorem Here we assume E is an event in a sample space S, and A1, A2, … An are mutually disjoint events whose union is S; that is, the events A1, A2, …, An form a partition of S. 40.16.
(Law of Total Probability) P(E) = P(A1)P(E|A1) + P(A2)P(E|A2) + … + P(An)P(E|An)
40.17. (Bayes’ Formula) For k = 1, 2, …, n, P(Ak|E) =
P( Ak )P(E | Ak ) P( Ak )P(E | Ak ) = P(E ) P( A1 )P(E | A1 ) + P( A2 )P(E | A2 ) + ⋅ ⋅ ⋅ + P( An )P(E | An )
EXAMPLE 40.3: Three machines, A, B, C, produce, respectively, 50%, 30%, and 20% of the total number of items in a factory. The percentages of defective output of these machines are, respectively, 3%, 4%, and 5%. An item is randomly selected.
(a) Find P(D), the probability the item is defective. (b) If the item is defective, find the probability it came from machine: (i) A, (ii) B, (iii) C. (a) By 40.16 (Total Probability Law), P(D) = P(A)P(D|A) + P(B)P(D|B) + P(C)P(D|C) = (0.50)(0.03) + (0.30)(0.04) + (0.20)(0.05) = 3.7% P( A)P( D | A) (0.50)(0.03) (b) By 40.17 (Bayes’ rule), (i) P(A|D) = = = 40.5%. Similarly, 0.037 P( D) P ( B ) P ( D | B) P(C )P( D | C ) = 32.5%; (iii) P(C|D) = = 27.0% (ii) P(B|D) = P( D) P( D)
PROBABILITY
221
Alternately, we may consider this problem as a two-step stochastic process with a probability tree diagram, as in Fig. 40-1(b). We find P(D) by adding the three probability paths to D: (0.50)(0.03) + (0.30)(0.04) + (0.20)(0.05) = 3.7% We find P(A|D) by dividing the top path to A and D by the sum of the three paths to D. (0.50)(0.03)/0.037 = 40.5% Similarly, we find P(B|D) = 32.5% and P(C|D) = 27.0%.
Independent Events DEFINITION 40.4:
Events A and B are independent if P(A ∩ B) = P(A)P(B).
40.18. The following are equivalent: (i) P(A ∩ B) = P(A)P(B), (ii) P(A|B) = P(A), (iii) P(B|A) = P(B). That is, events A and B are independent if the occurrence of one of them does not influence the occurrence of the other. EXAMPLE 40.4: Consider the following events for a family with children where we assume the sample space S is an equiprobable space:
E = {children of both sexes},
F = {at most one boy}
(a) Show that E and F are independent events if a family has three children. (b) Show that E and F are dependent events if a family has two children. (a) Here S = {bbb, bbg, bgb, bgg, gbb, gbg, ggb, ggg}. So: E = {bbg, bgb, bgg, gbb, gbg, ggb}, P(E) = 6/8 = 3/4, F = {bgg, gbg, ggb, ggg}, P(F) = 4/8 = 1/2 E ∩ F = {bgg, gbg, ggb}, P(E ∩ F) = 3/8 Therefore, P(E)P(F) = (3/4)(1/2) = 3/8 = P(E ∩ F). Hence, E and F are independent. (b) Here S = {bb, bg, gb, gg}. So: E = {bg, gb}, P(E) = 2/4 = 1/2, F = {bg, gb, gg}, P(F) = 3/4 E ∩ F = {bg, gb}, P(E ∩ F) = 2/4 = 1/2 Therefore, P(E)P(F) = (1/2)(3/4) = 3/8 ≠ P(E ∩ F). Hence, E and F are dependent. DEFINITION 40.5: and
For n > 2, the events A1, A2, …, An are independent if any proper subset of them is independent
P(A1 ∩ A2 ∩ … ∩ An) = P(A1)P(A2) … P(An) Observe that induction is used in this definition. DEFINITION 40.6: A collection {Aj | j ∈ J} of events is independent if, for any n > 0, the sets Aj , Aj , …, Aj are inn 1 2 dependent.
The concept of independent repeated trials, when S is a finite set, is formalized as follows.
PROBABILITY
222
DEFINITION 40.7: Let S be a finite probability space. The probability space of n independent trials or repeated trials, denoted by Sn, consists of ordered n-tuples (s1, s2, …, sn) of elements of S with the probability of an n-tuple defined by
P((s1, s2, …, sn)) = P(s1)P(s2) … P(sn) EXAMPLE 40.5: Suppose whenever horses a, b, c race together, their respective probabilities of winning are 20%, 30%, and 50%. That is, S = {a, b, c} with P(a) = 0.2, P(b) = 0.3, and P(c) = 0.5.
They race three times. Find the probability that (a) the same horse wins all three times (b) each horse wins once (a) Writing xyz for (x, y, z), we seek the probability of the event A = {aaa, bbb, ccc}. Here, P(aaa) = (0.2)3 = 0.008, P(bbb) = (0.3)3 = 0.027, P(ccc) = (0.5)3 = 0.125 Thus, P(A) = 0.008 + 0.027 + 0.125 = 0.160. (b) We seek the probability of the event B = {abc, acb, bac, bca, cab, cba}. Each element in B has the same probability (0.2)(0.3)(0.5) = 0.03. Thus, P(B) = 6(0.03) = 0.18.
41
RANDOM VARIABLES
Consider a probability space (S, C, P). DEFINITION 41.1. A random variable X on the sample space S is a function from S into the set R of real numbers such that the preimage of every interval of R is an event of S.
If S is a discrete sample space in which every subset of S is an event, then every real-valued function on S is a random variable. On the other hand, if S is uncountable, then certain real-valued functions on S may not be random variables. Let X be a random variable on S, where we let RX denote the range of X; that is, RX = {x | there exists s ∈S for which X(s) = x} There are two cases that we treat separately. (i) X is a discrete random variable; that is, RX is finite or countable. (ii) X is a continuous random variable; that is, RX is a continuum of numbers such as an interval or a union of intervals. Let X and Y be random variables on the same sample space S. Then, as usual, X + Y, X + k, kX, and XY (where k is a real number) are the functions on S defined as follows (where s is any point in S): (X + Y)(s) = X(s) + Y(s), (X + k)(s) = X(s) + k,
(kX)(s) = kX(s), (XY)(s) = X(s)Y(s).
More generally, for any polynomial, exponential, or continuous function h(t), we define h(X) to be the function on S defined by [h(X)](s) = h[X(s)] One can show that these are also random variables on S. The following short notation is used: P(X = xi) P(a ≤ X ≤ b μ X or E(X) or simply μ σ X or Var(X) or simply σ 2 σ X or simply σ
denotes the probability that X = xi. denotes the probability that X lies in the closed interval [a, b]. denotes the mean or expectation of X. denotes the variance of X. denotes the standard deviation of X.
2
Sometimes we let Y be a random variable such that Y = g(X), that is, where Y is some function of X.
Discrete Random Variables Here X is a random variable with only a finite or countable number of values, say RX = {x1, x2, x3, …}where, say, x1 < x2, < x3 < …. Then X induces a function f(x) on RX as follows: f(xi) = P(X = xi) = P({s ∈S | X(s) = xi}) The function f(x) has the following properties: (i) f(xi) ≥ 0
and
(ii) Σ i f(xi) = 1
Thus, f defines a probability function on the range RX of X. The pair (xi, f(xi)), usually given by a table, is called the probability distribution or probability mass function of X.
223
RANDOM VARIABLES
224
Mean 41.1.
μ X = E(X) = Σxif(xi) Here, Y = g(X).
41.2.
μY = E(Y) = Σg(xi) f(xi)
Variance and Standard Deviation 41.3.
σ X 2 = Var(X) = Σ(xi – m)2 f(xi) = E((X – m)2) Alternately, Var(X) = s 2 may be obtained as follows:
41.4. Var(X) = Sxi2f(xi) – m2 = E(X2) – m2 41.5. σX =
Var ( X ) =
E(X 2 ) − μ 2
REMARK: Both the variance Var(X) = s 2 and the standard deviation s measure the weighted spread of the values xi
about the mean m; however, the standard deviation has the same units as m. EXAMPLE 41.1:
Suppose X has the following probability distribution:
x
2
f(x)
0.1
4
6
8
0.2
0.3
0.4
Then: m = E(X) = Σxif(xi) = 2(0.1) + 4(0.2) + 6(0.3) + 8(0.4) = 6 E(X2) = Σxi2 f(xi) = 22(0.1) + 42(0.2) + 62(0.3) + 82(0.4) = 40 s 2 = Var(X) = E(X2) − m2 = 40 − 36 = 4 s=
Var ( X ) =
4 =2
Continuous Random Variable Here X is a random variable with a continuum number of values. Then X determines a function f(x), called the density function of X, such that (i) f(x) ≥ 0
and
(ii)
∫
∞ −∞
f(x) dx =
Furthermore, P(a ≤ X ≤ b) =
Mean 41.6.
μ X = E(X) =
∫
∞ −∞
xf(x) dx
Here, Y = g(X). 41.7.
μY = E(Y) =
∫
∞ −∞
g(x) f(x) dx
∫
b a
f(x) dx
∫
R
f ( x ) dx = 1
RANDOM VARIABLES
225
Variance and Standard Deviation ∞
41.8. σ X 2 = Var(X) = ∫ (x − m)2 f(x)dx = E((X − m)2) −∞ Alternately, Var(X) = s 2 may be obtained as follows: 41.9. Var(X) = 41.10. s X =
∫
∞ −∞
x2f(x)dx − m2 = E(X2) − m2
Var ( X ) =
E(X 2 ) − μ 2
Let X be the continuous random variable with the following density function:
EXAMPLE 41.2:
⎧(1 / 2) x f(x) = ⎨ ⎩ 0
if 0 ≤ x ≤ 2 elsewhere
Then: E(X) = ∫ E(X2) =
∞ −∞
∫
∞ −∞
2
xf(x) dx =
∫
x2f(x) dx =
⎡x3 ⎤ 1 2 4 x dx = ⎢ 6 ⎥ = 2 3 ⎣ ⎦0
2 0
∫
2 0
1 3 x dx = 2
s 2 = Var(X) = E(X2) − m2 = 2 − s=
Var ( X ) =
2
⎡x4 ⎤ ⎢⎣ 8 ⎥⎦ = 2 0
16 2 = 9 9
2 1 2 = 9 3
Cumulative Distribution Function The cumulative distribution function F(x) of a random variable X is the function F:R → R defined by 41.11. F(a) = P(X ≤ a) The function F is well-defined since the inverse of the interval (−∞, a] is an event. The function F(x) has the following properties: 41.12. 41.13.
F(a) ≤ F(b) whenever a ≤ b. lim F(x) = 0
x→−∞
and
lim F(x) = 1
x→+∞
That is, F(x) is monotonic, and the limit of F to the left is 0 and to the right is 1. If X is the discrete random variable with distribution f(x), then F(x) is the following step function: 41.14. F(x) =
∑ f(xi) xi ≤ x
If X is a continuous random variable, then the density funcion f(x) of X can be obtained from the cummulative distribution function F(x) by differentiation. That is, d F(x) = F′(x) dx Accordingly, for a continuous random variable X,
41.15. f(x) =
41.16.
F(x) =
∫
x −∞
f(t) dt
RANDOM VARIABLES
226
Standardized Random Variable The standardized random variable Z of a random variable X with mean m and standard deviation s > 0 is defined by X−μ σ Properties of such a standardized random variable Z follow:
41.17.
Z=
μ Z = E(Z) = 0
σZ = 1
and
Consider the random variable X in Example 41.1 where μ X = 6 and σ X = 2.
EXAMPLE 41.3:
The distribution of Z = (X – 6)/2 where f(z) = f(x) follows: Z
−2
−1
0
1
f(Z)
0.1
0.2
0.3
0.4
Then: E(Z) = Σ zif(zi) = (−2)(0.1) + (−1)(0.2) + 0(0.3) + 1(0.4) = 0 E(Z2) = Σ zi2 f(zi) = (−2)2(0.1) + (−1)2(0.2) + 02(0.3) + 12(0.4) = 1 Var(Z) = 1 − 02 = 1 and sZ = Var ( X ) = 1
Probability Distributions
∑
⎛ n⎞ t n-t ⎜⎝ t ⎟⎠ p q
41.18.
Binomial Distribution: Φ(x) =
41.19.
Poisson Distribution: Φ(x) =
41.20.
Hypergeometric Distribution: Φ(x) =
41.21.
Normal Distribution: Φ(x) =
41.22.
t≤x
∑
t≤x
λ t e −λ t!
1 2π
Student’s t Distribution: Φ(x) =
∫
∑
t≤x
x
⎛r⎞ ⎛ s ⎞ ⎜⎝ t⎟⎠ ⎜⎝ n − t⎟⎠ z ⎛r + s⎞ ⎜⎝ n ⎟⎠
e − t / 2 dt 2
−∞
⎛ n + 1⎞ Γ⎜ ⎝ 2 ⎟⎠ nπ Γ (n/2) 1
2 41.23. c (Chi Square) Distribution: Φ(x) =
2
n/2
∫
x −∞
⎛ t 2⎞ ⎜⎝1 + ⎟⎠ n
− ( n +1)/ 2
dt
x 1 t(n - 2)/2e-t/2 dt ∫ Γ (n/2) 0
⎛ n1 + n 2 ⎞ n /2 n Γ⎜ ⎟ n1 n 2 ⎝ 2 ⎠ F Distribution: Φ(x) = Γ (n 1 /2)Γ (n2 /2) 1
41.24.
p > 0, q > 0, p + q = 1
2
/2
∫
x 0
t (n
1
/ 2 ) −1
(n 2 + n 1 t ) − ( n + n 1
2
)/ 2
dt
Section XII: Numerical Methods
42
INTERPOLATION
Lagrange Interpolation Two-point formula 42.1.
x − x0 x − x1 + f ( x1 ) x 0 − x1 x1 − x 0
p ( x ) = f ( x0 )
where p (x) is a linear polynomial interpolating two points ( x 0 , f ( x 0 )), ( x1 , f ( x1 )), x 0 ≠ x1
General formula 42.2.
p ( x ) = f ( x 0 ) Ln , 0 ( x ) + f ( x1 ) Ln ,1 ( x ) + + f ( x n ) Ln , n ( x )
where Ln , k =
n
∏
i=0,i≠ k
x − xi x k − xi
and where p(x) is an nth-order polynomial interpolating n + 1 points ( x k , f ( x k )), k = 0, 1, … , n; and
xi ≠ x j for i ≠ j
Remainder formula Suppose f ( x ) ∈ Cn+1[a, b]. Then there is a ξ ( x ) ∈ (a, b) such that: 42.3.
f (x) = p (x) +
f n+1 (ξ ( x )) ( x − x 0 )( x − x1 ) ( x − x n ) (n + 1)!
Newton’s Interpolation First-order divided-difference formula 42.4.
f [ x 0 , x1 ] =
f ( x1 ) − f ( x 0 ) x1 − x 0
Two-point interpolatory formula 42.5.
p( x ) = f ( x 0 ) + f [ x 0 , x1 ]( x − x 0 )
where p(x) is a linear polynomial interpolating two points ( x 0 , f ( x 0 )), ( x1 , f ( x1 )), x 0 ≠ x1
227 Copyright © 2009, 1999, 1968 by The McGraw-Hill Companies, Inc. Click here for terms of use.
INTERPOLATION
228
Second-order divided-difference formula f [ x1 , x 2 ] − f [ x 0 , x1 ] 42.6. f [ x 0 , x1 , x 2 ] = x2 − x0 Three-point interpolatory formula 42.7.
p ( x ) = f ( x 0 ) + f [ x 0 , x1 ]( x − x 0 ) + f [ x 0 , x1 , x 2 ]( x − x 0 )( x − x1 )
where p(x) is a quadrant polynomial interpolating three points ( x 0 , f ( x 0 )), ( x1 , f ( x1 )), ( x 2 , f ( x3 ))
General kth-order divided-difference formula 42.8.
f [ x 0 , x1 ,… , x k } =
f [ x1 , x 2 ,… , x k ] − f [ x 0 , x1 ,… , x k −1 ] xk − x0
General interpolatory formula 42.9.
p ( x ) = f ( x 0 ) + f [ x 0 , x1 ]( x − x 0 ) + + f [ x 0 , x1 ,… , x n ]( x − x 0 )( x − x1 ) ( x − x n−1 )
where p(x) is an nth-order polynomial interpolating n + 1 points ( x k , f ( x k )), k = 0, 1,… , n; and
xi ≠ x j for i ≠ j
Remainder formula Suppose f ( x ) ∈ Cn+1[a, b]. Then there is a ξ ( x ) ∈ (a, b) such that 42.10.
f ( x ) = p( x ) +
f n+1 (ξ ( x )) ( x − x 0 )( x − x1 ) ( x − x n ) (n + 1)!
Newton’s Forward-Difference Formula First-order forward-difference at x0 42.11.
Δf ( x 0 ) = f ( x1 ) − f ( x 0 )
Second-order forward difference at x0 42.12.
Δ 2 f ( x 0 ) = Δf ( x1 ) − Δf ( x 0 )
General kth-order forward difference at x0 42.13.
Δ k f ( x 0 ) = Δ k −1 f ( x1 ) − Δ k −1 f ( x 0 )
Binomial coefficient 42.14.
⎛ s⎞ s(s − 1) (s − k + 1) ⎜⎝ k⎟⎠ = k!
Newton’s forward-difference formula ⎛ n⎞ p( x ) = ∑ ⎜ ⎟ Δ k f ( x 0 ) ⎝ k⎠ k =0 n
42.15.
where p(x) is an nth-order polynomial interpolating n + 1 equal spaced points ( x k , f ( x k )), x k = x 0 + kh k = 0, 1, … , n
INTERPOLATION
229
Newton’s Backward-Difference Formula First-order backward difference at xn 42.16.
∇f ( x n ) = f ( x n ) − f ( x n−1 )
Second-order backward difference at xn 42.17.
∇ 2 f ( x n ) = ∇f ( x n ) − ∇f ( x n−1 )
General kth-order backward difference at xn 42.18.
∇ k f ( x n ) = ∇ k −1 f ( x n ) − ∇ k −1 f ( x n−1 )
Newton’s backward-difference formula ⎛ − n⎞ p( x ) = ∑ (−1) k ⎜ ⎟ ∇ k f ( x n ) ⎝ k⎠ k =0 n
42.19.
where p(x) is an nth-order polynomial interpolating n + 1 equal spaced points ( x k , f ( x k )), x k = x 0 + kh
k = 0, 1, … , n
Hermite Interpolation Two-point basis polynomials 42.20.
x − x 0 ⎞ ( x − x1 )2 x − x1 ⎞ ( x − x 0 )2 ⎛ ⎛ H1,1 = ⎜1 − 2 H1, 0 = ⎜1 − 2 ⎟ 2 , x 0 − x1 ⎠ ( x 0 − x1 ) x1 − x 0 ⎟⎠ ( x1 − x 0 )2 ⎝ ⎝ ( x − x 0 )2 ( x − x1 )2 Hˆ 1, 0 = ( x − x 0 ) Hˆ 1,1 = ( x − x1 ) 2 , ( x 0 − x1 ) ( x1 − x 0 )2
Two-point interpolatory formula 42.21.
H 3 ( x ) = f ( x 0 ) H1, 0 + f ( x1 ) H1,1 + f ′( x 0 ) Hˆ 1, 0 + f ′( x1 ) Hˆ 1,1
where H3(x) is a third-order polynomial, agrees with f (x) and its first-order derivatives at two points, i.e., H 3 ( x 0 ) = f ( x 0 ), H 3′ ( x 0 ) = f ′( x 0 ),
H 3 ( x1 ) = f ( x1 ), H 3′ ( x1 ) = f ′( x1 )
General basis polynomials 42.22.
x − xj ⎞ 2 ⎛ H n , j = ⎜1 − 2 L ( x ), Hˆ n , j = ( x − x j ) L2n , j ( x ) Ln′, j ( x j )⎟⎠ n , j ⎝
where Ln , j =
n
∏
i=0,i≠ j
x − xi x j − xi
INTERPOLATION
230
General interpolatory formula 42.23.
n
n
j=0
j=0
H 2 n+1 ( x ) = ∑ f ( x j ) H n , j ( x ) + ∑ f ′( x j ) Hˆ n , j ( x )
where H 2 n+1 ( x ) is a (2n + 1)th-order polynomial, agrees with f(x) and its first order derivatives at n + 1 points, i.e., H 2 n+1 ( x k ) = f ( x k ), H 2′n+1 ( x k ) = f ′( x k )
Remainder formula Suppose f ( x ) ∈ C2 n+ 2 [a, b]. Then there is a ξ ( x ) ∈ (a, b) such that 42.24.
f ( x ) = H 2 n+1 ( x ) +
f 2 n+ 2 (ξ ( x )) ( x − x 0 )2 ( x − x1 )2 ( x − x n )2 (2n + 2)!
k = 0, 1, … , n
43
QUADRATURE
Trapezoidal Rule Trapezoidal rule 43.1.
∫
b a
f ( x ) dx ~
b−a [ f (a) + f (b)] 2
Composite trapezoidal rule
43.2.
∫
b a
f ( x ) dx ~
n −1 ⎞ h⎛ f (a) + 2∑ f (a + ih) + f (b)⎟ ⎜ 2⎝ ⎠ i =1
where h = (b − a)/n is the grid size.
Simpson’s Rule Simpson’s rule 43.3.
∫
b a
f ( x ) dx ~
b−a ⎡ ⎤ ⎛ a + b⎞ f (a) + 4 f + f (b)⎥ 6 ⎢⎣ ⎝ 2 ⎠ ⎦
Composite Simpson’s rule 43.4.
∫
b a
f ( x ) dx ~
n/2 n/2 ⎞ h⎛ 2 4 f ( x ) + f ( x ) + f ( x 2i −1 ) + f ( x n )⎟ ∑ ∑ 0 2 i − 2 3 ⎜⎝ ⎠ i=2 i =1
where n even, h = (b − a)/n, xi = a + ih, i = 0, 1, … , n.
Midpoint Rule Midpoint rule 43.5.
∫
b a
f ( x ) dx ~ (b − a) f
⎛ a + b⎞ ⎝ 2 ⎠
Composite midpoint rule 43.6.
∫
b a
n/2
f ( x ) dx ~ 2h∑ f ( x 2i ) i=0
where n even, h = (b − a)/(n + 2), xi = a + (i − 1)h, i = −1, 0, … , n + 1.
231
QUADRATURE
232
Gaussian Quadrature Formula Legendre polynomial 43.7.
Pn ( x ) =
1 dn [( x 2 − 1)n ] 2n n ! dx n
Abscissa points and weight formulas The abscissa points x k( n ) and weight coefficient ω k( n ) are defined as follows: 43.8.
x k( n ) = the kth zero of the Legendre polynomial Pn(x)
43.9.
ω k( n ) =
2Pn′( x k( n ) )2 1 − x k( n ) 2
Tables for Gauss-Legendre abscissas and weights appear in Fig. 43-1. Gauss-Legendre formula in interval (–1, 1) 43.10.
∫
1 −1
n
f ( x ) dx = ∑ ω k( n ) f ( x k( n ) ) + Rn k =1
Gauss-Legendre formula in general interval (a, b) 43.11.
∫
b a
f ( x ) dx =
b − a n (n) ⎛ a + b b − a⎞ ωk f + x k( n ) + Rn 2 ∑ 2 2 ⎠ ⎝ k =1
Remainder formula 43.12.
Rn =
(b − a)2 n+1 (n !)4 ( 2 n ) f (ξ ) (2n + 1)[(2n)!]3
for some a < ξ < b.
Fig. 43-1
44
SOLUTION of NONLINEAR EQUATIONS
Here we give methods to solve nonlinear equations which come in two forms: 44.1.
Nonlinear equation: f (x) = 0
44.2. Fixed point nonlinear equation: x = g(x) One can change from 44.1 to 44.2 or from 44.2 to 44.1 by settting: g( x ) = f ( x ) + x or
f ( x ) = g( x ) − x
Since the methods are iterative, there are two types of error estimates: 44.3.
| f ( x n ) | < or | x n+1 − x n | <
for some preassigned > 0.
Bisection Method The following theorem applies: Intermediate Value Theorem: Suppose f is continuous on an interval [a, b] and f (a) f (b) < 0. Then there is a root x* to f (x) = 0 in (a, b). The bisection method approximates one such solution x*. 44.4.
Bisection method: Initial step: Set a0 = a and b0 = b. Repetitive step: (a) Set cn = (an + bn )/2. (b) If f (an ) f (cn ) < 0, then set an+1 = an and bn+1 = cn ; else set an+1 = cn and bn+1 = bn .
Newton’s Method Newton method 44.5.
x n+1 = x n −
f ( xn ) f ′( x n )
Quadratic convergence 44.6.
| x n+1 − x ∗ | f ′′( x ∗ ) = ∗ 2 n→∞ | x − x | 2( f ′( x ∗ ))2 n
lim
where x* is a root of the nonlinear equation 44.1.
233
SOLUTION OF NONLINEAR EQUATIONS
234
Secant Method Secant method 44.7.
x n+1 = x n −
( x n − x n−1 ) f ( x n ) f ( x n ) − f ( x n−1 )
Rate of convergence 44.8.
| x n+1 − x ∗ | f ′′( x ∗ ) = n→∞ | x − x ∗ || x − x ∗ | 2( f ′( x ∗ ))2 n n −1
lim
where x* is a root of the nonlinear equation 44.1.
Fixed-Point Iteration The following definition and theorem apply: Definition: A function g from (a, b) to (a, b) is called a contraction mapping if | g ( x ) − g ( y) | L | x − y |
for any x , y ∈ (a, b)
where L < 1 is a positive constant. Fixed-point theorem: Suppose that g is a contraction mapping on (a, b). Then g has a unique fixed point in (a, b). Given such a contraction mapping g, the following method may be used. Fixed-point iteration 44.9.
x n+1 = g( x n )
45
NUMERICAL METHODS for ORDINARY DIFFERENTIAL EQUATIONS
Here we give methods to solve the following initial-value problem of an ordinary differential equation:
45.1.
⎧⎪dx = f ( x , t ) ⎨ dt ⎩⎪x (t0 ) = x 0
The methods will use a computational grid: 45.2.
tn = t0 + nh
where h is the grid size.
First-Order Methods Forward Euler method (first-order explicit method) 45.3.
x (t + h) = x (t ) + hf ( x (t ), t )
Backward Euler method (first-order implicit method) 45.4.
x (t + h) = x (t ) + hf ( x (t + h), t + h)
Second-Order Methods Mid-point rule (second-order explicit method)
45.5.
⎧ * h ⎪⎪x = x (t ) + 2 f ( x (t ), t ) ⎨ ⎛ h⎞ ⎪x (t + h) = x (t ) + hf ⎜ x * , t + ⎟ 2⎠ ⎝ ⎪⎩
Trapezoidal rule (second-order implicit method) 45.6.
h x (t + h) = x (t ) + { f ( x (t ), t ) + f ( x (t + h), t + h)} 2
Heun’s method (second-order explicit method)
45.7.
⎧x * = x (t ) + hf ( x (t ), t ) ⎪ h ⎨ * ⎪⎩x (t + h) = x (t ) + 2 { f ( x (t ), t ) + f ( x , t + h)}
235
NUMERICAL METHODS FOR ORDINARY DIFFERENTIAL EQUATIONS
236
Single-Stage High-Order Methods Fourth-order Runge–Kutta method (fourth-order explicit method) 45.8.
x (t + h) = x (t ) +
1 (F + 2F2 + 2F3 + F4 ) 6 1
where F ⎛ F1 = hf ( x , t ), F2 = hf ⎜ x + 1 , t + 2 ⎝
F h⎞ ⎛ , F3 = hf ⎜ x + 2 , t + 2⎟⎠ 2 ⎝
h⎞ , F4 = hf ( x + F3 , t + h) 2⎟⎠
Multi-Step High-Order Methods Adams-Bashforth two-step method 45.9.
x (t + h) = x (t ) + h
1 ⎛3 f ( x (t ), t ) − f ( x (t − h), t − h)⎞ 2 ⎝2 ⎠
Adams-Bashforth three-step method 45.10.
x (t + h) = x (t ) + h
5 4 ⎛ 23 f ( x (t − 2h), t − 2h)⎞ f ( x (t ), t ) − f ( x (t − h), t − h) + 12 3 ⎝ 12 ⎠
Adams-Bashforth four-step method 45.11.
⎛ 55 ⎞ 59 37 9 x (t + h) = x (t ) + h ⎜ f ( x (t ), t ) − f ( x (t − h), t − h) + f ( x (t − 2h), t − 2h) − f ( x (t − 3h), t − 3h)⎟ 24 24 24 ⎝ 24 ⎠
Milne’s method 45.12.
x (t + h) = x (t − 3h) + h
8 4 ⎛8 f ( x (t ), t ) − f ( x (t − h), t − h) + f ( x (t − 2h), t − 2h)⎞ 3 3 ⎝3 ⎠
Adams-Moulton two-step method 45.13.
x (t + h) = x (t ) + h
2 1 ⎛5 f ( x (t + h), t + h) + f ( x (t ), t ) − f ( x (t − h), t − h)⎞ 3 12 ⎝ 12 ⎠
Adams-Moulton three-step method
45.14.
⎛3 ⎞ 5 1 19 x (t + h) = x (t ) + h ⎜ f ( x (t + h), t + h) + f ( x (t ), t ) − f ( x (t − h), t − h) + f ( x (t − 2h), t − 3h)⎟ 24 24 24 ⎝8 ⎠
46
NUMERICAL METHODS for PARTIAL DIFFERENTIAL EQUATIONS
Finite-Difference Method for Poisson Equation The following is the Poisson equation in a domain (a, b) × (c, d): 46.1.
∇ 2u = f ,
∇2 =
∂2 ∂2 2 + ∂x ∂y 2
Boundary condition: 46.2.
u ( x , y) = g ( x , y)
for x = a, b
or
y = c, d
Computation grid: 46.3.
xi = a + iΔx
for i = 0, 1, … , n
y j = c + jΔy
for j = 0, 1, … , m
where Δx = (b − a)/n and Δy = (d − c)/m are grid sizes for x and y variables, respectively. Second-order difference approximation 46.4.
( Dx2 + Dy2 )u( xi , y j ) = f ( xi , y j )
where Dx2 u( xi , y j ) =
u( xi +1 , y j ) − 2u( xi , y j ) + u( xi −1 , y j ) Δx 2
Dy2 u( xi , y j ) =
u( xi , y j +1 ) − 2u( xi , y j ) + u( xi , y j −1 ) Δy 2
Computational boundary condition 46.5.
u( x 0 , y j ) = g(a, y j ),
u( x n , y j ) = g(b, y j )
for j = 1, 2, … , m
u( xi , y0 ) = g( xi , c),
u( xi , ym ) = g( xi , d )
for i = 1, 2, … , n
Finite-Difference Method for Heat Equation The following is the heat equation in a domain (a, b) × (c, d ) × (0, T ) : 46.6.
∂u = ∇ 2u ∂t
237
NUMERICAL METHODS FOR PARTIAL DIFFERENTIAL EQUATIONS
238
Boundary condition: 46.7.
u ( x , y , t ) = g ( x , y)
for x = a, b
or
y = c, d
Initial condition: 46.8.
u( x , y, 0) = u0 ( x , y)
Computational grid: 46.9.
xi = a + iΔx
for i = 0,1, … , n
y j = c + jΔy
for j = 0,1, … , m
t k = kΔt
for k = 0,1, … ,
where Δx = (b − a)/n, Δy = (d − c)/m, and Δt are grid sizes for x, y and t variables, respectively. Computational boundary condition 46.10.
u( x 0 , y j ) = g(a, y j ), u( x n , y j ) = g(b, y j )
for j = 1, 2, … , m
u( xi , y0 ) = g( xi , c), u( xi , ym ) = g( xi , d )
for i = 1, 2, … , n
Computational initial condition 46.11.
u( xi , y j , 0) = u0 ( xi , y j )
for i = 1, 2, … , n; j = 0, 1, … , m
Forward Euler method with stability condition 46.12.
u( xi , y j , t k +1 ) = u( xi , y j , t k ) + Δt ( Dx2 + Dy2 )u( xi , y j , t k )
46.13.
2Δt 2Δt + 1 Δx 2 Δy 2
Backward Euler method (unconditional stable) 46.14.
u( xi , y j , t k +1 ) = u( xi , y j , t k ) + Δt ( Dx2 + Dy2 )u( xi , y j , t k +1 )
Crank-Nicholson method (unconditional stable) 46.15.
u( xi , y j , t k +1 ) = u( xi , y j , t k ) + Δt ( Dx2 + Dy2 ){u( xi , y j , t k ) + u( xi , y j , t k +1 )}/2
Finite-Difference Method for Wave Equation The following is a wave equation in a domain (a, b) × (c, d ) × (0, T ): 46.16.
∂2u = A2 ∇ 2 u ∂t 2
where A is a constant representing the speed of the wave.
NUMERICAL METHODS FOR PARTIAL DIFFERENTIAL EQUATIONS
Boundary condition: 46.17.
u ( x , y, t ) = g ( x , y)
for x = a, b or y = c, d
Initial condition: 46.18.
u( x , y, 0) = u0 ( x , y),
∂u u( x , y, 0) = u1 ( x , y) ∂t
Computational grids: 46.19.
xi = a + iΔx
for i = 0, 1, … , n
y j = c + jΔy
for j = 0, 1, … , m
t k = kΔt
for k = −1, 0, 1, …
where Δx = (b − a)/n, Δy = (d − c)/m, and Δt are the grid sizes for x, y, and t variables, respectively. A second-order finite-difference approximation 46.20.
u( xi , y j , t k +1 ) = 2u( xi , y j , t k ) − u( xi , y j , t k −1 ) + Δt 2 A2 ( Dx2 + Dy2 )u( xi , y j , t k )
Computational boundary condition 46.21.
u( x 0 , y j ) = g(a, y j ), u( x n , y j ) = g(b, y j )
for j = 1, 2, … , m
u( xi , y0 ) = g( xi , c), u( xi , ym ) = g( xi , d )
for i = 1, 2, … , n
Computational initial condition 46.22.
u( xi , y j , t0 ) = u0 ( xi , y j )
for i = 1, 2, … , n; j = 0, 1, … , m
u( xi , y j , t −1 ) = u0 ( xi , y j ) + Δt 2 u1 ( xi , y j )
Stability condition 46.23.
Δt A min(Δx , Δx )
for i = 1, 2, … , n; j = 0, 1, … , m
239
ITERATION METHODS for LINEAR SYSTEMS
47
Iteration Methods for Poisson Equation The finite-difference approximation to the Poisson equation follows: 47.1.
⎧ui +1, j + ui −1, j + ui , j +1 + ui , j −1 − 4ui , j = fi , j ⎪⎪ for j = 1, 2, … , n − 1 ⎨u0 , j = un , j = 0 ⎪ for i = 1, 2, … , n − 1 ⎪⎩ui ,0 = ui ,n = 0
for i, j = 1, 2, … , n − 1
Three iteration methods for solving the system follow: Jacobi method 47.2.
uik,+j 1 =
1 k (u + uik−1, j + uik, j +1 + uik, j −1 − fi , j ) 4 i +1, j
Gauss-Seidel method 47.3.
uik,+j 1 =
1 k (u + uik−+11, j + uik, j +1 + uik,+j −11 − fi , j ) 4 i +1, j
Successive-overrelaxation (SOR) method 47.4.
1 k ⎧ * * k * ⎪ui , j = (ui +1, j + ui −1, j + ui , j +1 + ui , j −1 − fi , j ) 4 ⎨ ⎪⎩uik,+j 1 = (1 − ω )uik, j + ω ui*, j
Iteration Methods for General Linear Systems Consider the linear system 47.5. Ax = b where A is an n × n matrix and x and b are n-vectors. We assume the coefficient matrix A is partitioned as follows: 47.6. A = D – L – U where D = diag (A), L is the negative of the strictly lower triangular part of A, and U is the negative of the strictly upper triangular part of A.
240
ITERATION METHODS FOR LINEAR SYSTEMS
Four iteration methods for solving the system follow: Richardson method 47.7.
x k +1 = (I − A) x k + b
Jacobi method 47.8.
Dx k +1 = ( L + U ) x k + b
Gauss-Seidel method 47.9.
( D − L ) x k +1 = Ux k + b
Successive-overrelaxation (SOR) method 47.10.
( D − ω L ) x k +1 = ω (Ux k + b) + (1 − ω ) Dx k
241
This page intentionally left blank
PART B
TABLES
Copyright © 2009, 1999, 1968 by The McGraw-Hill Companies, Inc. Click here for terms of use.
This page intentionally left blank
Section I: Logarithmic, Trigonometric, Exponential Functions
1
FOUR PLACE COMMON LOGARITHMS log 10 N or log N
245 Copyright © 2009, 1999, 1968 by The McGraw-Hill Companies, Inc. Click here for terms of use.
1
246
FOUR PLACE COMMON LOGARITHMS log 10 N or log N (Continued)
2
Sin x (x in degrees and minutes)
247
3
248
Cos x (x in degrees and minutes)
4
Tan x (x in degrees and minutes)
249
5
250
CONVERSION OF RADIANS TO DEGREES, MINUTES, AND SECONDS OR FRACTIONS OF DEGREES
6
CONVERSION OF DEGREES, MINUTES, AND SECONDS TO RADIANS
251
7
NATURAL OR NAPIERIAN LOGARITHMS log e x or ln x
ln 10 = 2.30259 2 ln 10 = 4.60517 3 ln 10 = 6.90776
252
4 ln 10 = 9.21034 5 ln 10 = 11.51293 6 ln 10 = 13.81551
7 ln 10 = 16.11810 8 ln 10 = 18.42068 9 ln 10 = 20.72327
7
NATURAL OR NAPIERIAN LOGARITHMS log e x or ln x(Continued)
253
8
254
EXPONENTIAL FUNCTIONS ex
9
EXPONENTIAL FUNCTIONS e –x
255
10
256
EXPONENTIAL, SINE, AND COSINE INTEGRALS Ei( x ) =
∫
∞ −u x
e du, Si( x ) = u
∫
x 0
sin u du, Ci( x ) = u
∫
∞ x
cos u du u
Section II: Factorial and Gamma Function, Binomial Coefficients
11
FACTORIAL n n ! = 1 i 2 i 3 ii n
257 Copyright © 2009, 1999, 1968 by The McGraw-Hill Companies, Inc. Click here for terms of use.
12
258
GAMMA FUNCTION Γ( x ) =
∫
∞
x
t x −1e − t dt
for 1 x 2
[For other values use the formula Γ(x + 1) = x Γ(x)]
13
BINOMIAL COEFFICIENTS n! n(n − 1) (n − k + 1) ⎛ n ⎞ ⎛ n⎞ =⎜ , 0! = 1 ⎜⎝ k⎟⎠ = k !(n − k )! = k! ⎝ n − k⎟⎠
Note that each number is the sum of two numbers in the row above; one of these numbers is in the same column and the other is in the preceding column (e.g., 56 = 35 + 21). The arrangement is often called Pascal’s triangle (see 3.6, page 8).
259
13
BINOMIAL COEFFICIENTS n! n(n − 1) (n − k + 1) ⎛ n ⎞ ⎛ n⎞ =⎜ , 0 ! = 1 (Continued) ⎜⎝ k⎟⎠ = k !(n − k )! = k! ⎝ n − k⎟⎠
⎛ n⎞ ⎛ n ⎞ . For k > 15 use the fact that ⎜ ⎟ = ⎜ ⎝ k⎠ ⎝ n − k⎟⎠
260
Section III: Bessel Functions
14
15
BESSEL FUNCTIONS J 0 (x)
BESSEL FUNCTIONS J 1 (x)
261 Copyright © 2009, 1999, 1968 by The McGraw-Hill Companies, Inc. Click here for terms of use.
16
17
262
BESSEL FUNCTIONS Y 0 (x)
BESSEL FUNCTIONS Y 1 (x)
18
19
BESSEL FUNCTIONS I 0 (x)
BESSEL FUNCTIONS I 1 (x)
263
20
21
264
BESSEL FUNCTIONS K 0 (x)
BESSEL FUNCTIONS K 1 (x)
22
23
BESSEL FUNCTIONS Ber(x)
BESSEL FUNCTIONS Bei(x)
265
24
25
266
BESSEL FUNCTIONS Ker(x)
BESSEL FUNCTIONS Kei(x)
26
VALUES FOR APPROXIMATE ZEROS OF BESSEL FUNCTIONS
The following table lists the first few positive roots of various equations. Note that for all cases listed the successive large roots differ approximately by π = 3.14159. . . .
267
Section IV: Legendre Polynomials
27
LEGENDRE POLYNOMIALS P n (x) [P 0 (x)=1, P 1 (x)=x]
268 Copyright © 2009, 1999, 1968 by The McGraw-Hill Companies, Inc. Click here for terms of use.
28
LEGENDRE POLYNOMIALS P n (cos ) [P 0 (cos )=1]
269
Section V: Elliptic Integrals
29
COMPLETE ELLIPTIC INTEGRALS OF FIRST AND SECOND KINDS K=
∫
π /2 0
dθ
1 − k sin θ 2
2
, E =∫
π /2 0
1 − k 2 sin 2 θ dθ , k = sin ψ
270 Copyright © 2009, 1999, 1968 by The McGraw-Hill Companies, Inc. Click here for terms of use.
30
31
INCOMPLETE ELLIPTIC INTEGRAL OF THE FIRST KIND F (k , φ ) =
∫
φ 0
dθ , k = sin ψ 1 − k 2 sin 2 θ
INCOMPLETE ELLIPTIC INTEGRAL OF THE SECOND KIND E (k , φ ) =
∫
φ 0
1 − k 2 sin 2 θ dθ , k = sin ψ
271
Section VI: Financial Tables
32
COMPOUND AMOUNT : (1 + r) n If a principal P is deposited at interest rate r (in decimals) compounded annually, then at the end of n years the accumulated amount A = P(1 + r)n.
272 Copyright © 2009, 1999, 1968 by The McGraw-Hill Companies, Inc. Click here for terms of use.
33
PRESENT VALUE OF AN AMOUNT : (1 ⴙ r) ⴚn The present value P which will amount to A in n years at an interest rate of r (in decimals) compounded annually is P = A(1 + r)n.
273
AMOUNT OF AN ANNUITY :
34
274
(1 + r ) n - 1 r
If a principal P is deposited at the end of each year at interest rate r (in decimals) compounded annually, then at the end of n years the accumulated amount is ⎡(1 − r )n − 1⎤ . The process is often called an annuity. P⎢ r ⎣ ⎦⎥
PRESENT VALUE OF AN ANNUITY :
35
1 - (1 + r ) - n r
An annuity in which the yearly payment at the end of each of n years is A at an interest rate r (in decimals) compounded annually has present value ⎡1 − (1 + r ) − n ⎤ . A⎢ ⎥⎦ r ⎣
275
Section VII: Probability and Statistics
36 NOTE:
AREAS UNDER THE STANDARD NORMAL CURVE from −∞ to x x 1 Φ( x ) = e − t / 2 dt ∫ −∞ 2π 2
erf (x) = 2Φ(x 2 ) − 1
276 Copyright © 2009, 1999, 1968 by The McGraw-Hill Companies, Inc. Click here for terms of use.
37
ORDINATES OF THE STANDARD NORMAL CURVE y=
1 −x e 2π
2
/2
277
38
278
PERCENTILE VALUES (t p ) FOR STUDENT'S t DISTRIBUTION with n degrees of freedom (shaded area = p)
39
PERCENTILE VALUES ( 2p ) FOR 2 (CHI-SQUARE) DISTRIBUTION with n degrees of freedom (shaded area = p)
279
40
280
95th PERCENTILE VALUES FOR THE F DISTRIBUTION n1 = degrees of freedom for numerator n2 = degrees of freedom for denominator (shaded area = .95)
41
99th PERCENTILE VALUES FOR THE F DISTRIBUTION n1 = degrees of freedom for numerator n2 = degrees of freedom for denominator (shaded area = .99)
281
42
282
RANDOM NUMBERS
Index of Special Symbols and Notations The following list show special symbols and notations together with pages on which they are defined or first appear. Cases where a symbol has more than one meaning will be clear from the context.
Symbols Bern(x), Bein(x) B(m, n) Bb C(x) Ci(x) e1, e2, e3 erf(x) erfc(x) E = E(k, p /2) E = E(k, f) Ei(x) En E(X) f [x0, x1, ..., xk] F(a), F(x) F(a, b; c; x) F(k, f ) g, g−1 G. M. h1, h2, h3 Hn(x) Hn(1)(x), Hn(2)(x) H. M. i, j, k In(x) Jn(x) K = F(k, p /2) Kern(x), Kein(x) Kn(x) ln x or loge x log x or log10 x Ln(x) Lnm(x) l, l−1 M.D. P(A/E) Pn(x) Pnm(x) QU, M, QL Qn(x) Qnm(x) r R.M.S. s s2 sxy Si(x) S(x) Tn(x)
Ber and Bei functions, 157 beta function, 152 Bernoulli numbers, 142 Fresnel cosine integral, 204 cosine integral, 204 unit vectors in curvilinear coordinates, 127 error function, 203 complementary error function, 203 complete elliptic integral of the second kind, 198 incomplete elliptic integral of the second kind, 198 exponential integral, 203 Euler number, 142 mean or expectation of random variable X, 223 divided distance formula, 287, 288 cumulative distribution function, 209 hypergeometric function, 178 incomplete elliptic integral of the first kind, 198 Fourier transform and inverse Fourier transform, 194 geometric mean, 209 scale factors in curvilinear coordinates, 127 Hermite polynomial, 169 Hankel functions of the first and second kind, 155 harmonic mean, 210 unit vectors in rectangular coordinates, 120 modified Bessel function of the first kind, 155 Bessel function of the first kind, 153 complete elliptic integral of the first kind, 198 Ker and Kei functions, 158 modified Bessel function of the second kind, 156 natural logarithm of x, 53 common logarithm of x, 53 Laguerre polynomials, 171 associated Laguerre polynomials, 173 Laplace transform and inverse Laplace transform, 180 mean deviation conditional probability of A given E, 219 Legendre polynomials, 164 associated Legendre polynomials, 173 quartiles, 211 Legendre functions of second kind, 167 associated Legendre functions of second kind, 168 sample correlation coefficient, 213 root-mean-square, 211 sample standard deviation, 208 sample variance, 210 sample covariance, 213 Sine integral, 203 Fresnel sine integral, 204 Chebyshev polynomials of first kind, 175
283 Copyright © 2009, 1999, 1968 by The McGraw-Hill Companies, Inc. Click here for terms of use.
INDEX OF SPECIAL SYMBOLS AND NOTATIONS
284
Un(x) Var(X) x, x xk(n) Yn(x) Z
Chebyshev polynomials of second kind, 176 variance of random variable X, 224 sample mean, grand mean, 208, 209 kth zero of Legendre polynomial Pn(x), 232 Bessel function of second kind, 153 standardized random variable, 226
Greek Symbols ar g Γ(x) ζ(x) m q
rth moment in standard units, 212 Euler’s constant, 4 gamma function, 149 Rieman zeta function, 204 population mean, 208 coordinate: cylindrical 37, polar, 11, 24; spherical, 38
p f Φ (p) Φ (x) s s2
pi, 3 spherical coordinate, 38 1 1 1 sum 1 + + + + , Φ(0) = 0, 154 p 2 3 probability distribution function, 226 population standard deviation, 223 population variance, 223
Notations A~B |A|
A is asymptotic to B or A/B approaches 1, 151 ⎧ A if A ≥ 0 absolute value of A = ⎨ ⎩− A if A < 0
n!
factorial n, 7
⎛ n⎞ ⎜⎝ k⎟⎠ dy ⎫ = f ( x ) ⎪ dx ⎬ d2y y = 2 = f ( x ), etc.⎪ dx ⎭ p d Dp = p dx ∂f ∂f ∂2 f , , , etc. ∂x ∂x ∂x ∂y ∂( x , y, z ) ∂(u1 , u2 , u3 )
binomial coefficients, 8
y =
∫ f ( x )dx b
∫a f ( x )dx ∫C A i dr
AiB A×B ∇ ∇2 = ∇ i ∇ ∇4 = ∇2 (∇2)
derivatives of y or f(x) with respect to x, 62
pth derivative with respect to x, 64 partial derivatives, 65 Jacobian, 128 indefinite integral, 67 definite integral, 108 line integral of A along C, 124 dot product of A and B, 120 cross product of A and B, 121 del operator, 122 Laplacian operator, 123 biharmonic operator, 123
Index Adams-Bashforth methods, 236 Adams-Moulton methods, 236 Addition formula: Bessel functions, 163 Hermite polynomials, 170 Addition rule (probability) 208 Addition of vectors, 119 Algebra of sets, 217 Algebraic equations, solutions of, 13 Alphabet, Greek, 3 Analytic geometry, plane, 22–33 solid, 34–40 Annuity table, 274 Anti-derivative, 67 Anti-logarithms, 53 Arithmetic: mean, 208 series, 134 Arithmetic-geometric series, 134 Associated Laguerre polynomials, 173 (See also Laguerre polynomials) Associated Legendre functions, 164 (See also Legendre functions) of the first kind, 168 of the second kind, 168 Asymptotic expansions or formulas: Bernoulli numbers, 143 Bessel functions, 160 Backward difference formulas, 228 Her and Bei functions, 157 Bayes formula, 220 Bernoulli numbers, 142 asymptotic formula, 143 series, 143 Bernoulli’s differential equation, 116 Bessel functions, 153–164 graphs, 159 integral representation, 161 modified, 155 recurrence formulas, 154, 157 series, orthogonal, 161 tables, 261–267 Bessel’s differential equation, 118, 153 general solution, 154 modified differential equation, 155 Best fit, line of, 214 Beta function, 152 Biharmonic operator, 123 Binomial: coefficients, 7, 228, 259 distribution, 226 formula, 7 series, 136 Bipolar coordinates, 131 Bisection method, 223 Bivariate data, 212 Carioid, 29 Cassini, ovals of, 32 Catalan’s constant, 200 Catenary, 29 Cauchy or Euler differential equation, 117
Cauchy’s form of remainder in Taylor series, 134 Cauchy-Schwarz inequality, 205 for integrals, 206 Central tendency, 208 Chain rule for derivatives, 67 Chebyshev polynomials, 175 of the first kind, 175 of the second kind, 176 recurrence formula, 175 Chebyshev’s differential equation, 175 general solution, 177 Chebyshev’s inequality, 206 Chi-square distribution, 226 table of values, 279 Circle, 17, 25 Coefficient: of excess (kurtosis), 212 of skewness, 212 Coefficients: binomial, 7 multinomial, 9 Complementary error function, 203 Complex: conjugate, 10 numbers, 10 logarithm of, 55 plane, 10 Components of a vector, 120 Compound amount, 262 Confocal: ellipsdoidal coordinates, 133 paraboloidal coordinates, 133 Conical coordinates, 129 Conics, 25 (See also Ellipse, Parabola, Hyperbola) Conjugate, complex, 10 Constant of integration, 67 Constants, 3 series of, 134 Continuous random variable, 224 Convergence, interval of, 138. Conversion factors, 15 Convolution theorem, Fourier transform, 194 Coordinates, 127 bipolar, 131 confocal ellipsoidal, 133 confocal paraboloidal, 133 conical, 132 curvilinear, 127 cylindrical, 129 elliptic cylindrical, 130 oblate spheroidal, 131 paraboloidal, 130 prolate spheroidal, 131 spherical, 129 toroidal, 132 Correlation coefficient, 213 Cosine, 43 graph of, 46 table of values, 245 Cosine integral, 203, 256 Cosines, law of, 51 Covariance, 213 Cross or vector product, 121 Cubic equation, solution of, 13
285 Copyright © 2009, 1999, 1968 by The McGraw-Hill Companies, Inc. Click here for terms of use.
INDEX
286
Cumulative distribution function, 225 Curl, 123 Curve fitting, 215 Curvilinear coordinates, 134 Cycloid, 28 Cylindrical coordinates, 37, 129 Definite integrals, 108–116 approximate formula, 109 definition of, 108 Degrees, conversion to radians, 251 Del operator, 122 DeMoivre’s theorem, 11 Derivatives, 62–66 chain rule for, 62 higher, 64 Leibniz’s rule, 64 of vectors, 122 Deviation: mean, 210 standard, 210 Differential equations, numerical methods for solution: ordinary, 235–236 partial, 237–240 Differentials, 65, 66 Differentiation, 62–66 (See also Derivatives) Direction numbers, 34 cosines, 34 Discrete random variable, 223 Distributions, probability, 226 Divergence, 122, 128 theorem, 126 Divided-difference formula (general), 228 Dot or scalar product, 120 Double integrals, 125 Eccentricity, 25 Ellipse, 18, 25 Ellipsoid, 39 Elliptic cylinder, 41 Elliptic cylindrical coordinates, 130 Elliptic functions, 198–202 Jacobi’s, 199 series expansion, 200 Elliptic integrals, 198–199 table of values, 270–271 Epicycloid, 30 Equality of vectors, 119 Equations, algebraic, 13 Error functions, 203 Euler: constant, 4 differential equation, 117 methods, 235 numbers, 142 Euler-Maclaurin summation formula, 137 Exact differential equation, 116 Excess, coefficient of kurtosis, 212 Exponential curve (least-squares), 215 Exponential function, 53–54 series for, 139 table of values, 254–255 Exponential integral, 203, 256 Exponents, 53 F distribution, 226 table of values, 280–281 Factorial n, 7 table of values, 257 Factors, special, 5 Financial tables, 272–275
Finite-difference methods for solution of: heat equation, 237 Poisson equation, 237 wave equation, 238 First-order divided-difference formula, 227 Five number summary [L, QL, M, QH, H ], 211 Fixed-point iteration, 234 Folium of Descartes, 31 Forward difference formulas, 228 Fourier series, 144–146 Fourier transform, 193 convolution of, 194 cosine, 194, 197 Parseval’s identity for, 193 sine, 194, 196 tables, 195–199 Fourier’s integral theorem, 193 Fresnel sine and cosine integral, 204 Frullani’s integral, 115 Gamma function, 149, 150 relation to beta function, 152 table of values, 258 Gauss’ theorem, 126 Gauss-Legendre formula, 232 Gauss-Seidel method, 230 Gaussian quadrature formula, 231 Generating functions, 157, 165, 168, 169, 171, 173, 175, 176 Geometric: mean (G.M.), 209 series, 134 Geometry, 16–21 analytic, 22–40 Gradient, 122, 128 Grand mean, 209 Greek alphabet, 3 Green’s theorem, 126 Griggsian logarithms, 53 Half angle formulas, 48 Half rectified sine wave function, 191 Hankel functions, 155 Harmonic mean, 209 Heat equation, 237 Heaviside’s unit function, 192 Hermite: interpolation,229 polynomials, 169–170 Hermite’s differential equation, 169 Heun’s method, 235 Holder’s inequality, 205 for integrals, 206 Homogeneous differential equation, 116 linear second order, 117 Hyperbola, 25 Hyperbolic functions, 56–61 graphs of, 59 inverse, 59–61 series for, 140 Hyperboloid, 39 Hypergeometric: differential equation, 178 distribution, 226 functions, 178 Hypocycloid, 28, 30 Imaginary part of a complex number, 10 Indefinite integrals, 67–107 definition of, 67 tables of, 71–107 transformation of, 69 Independent events, 221
INDEX
Inequalities, 205 Infinite products, 207 Integral calculus, fundamental theorem, 108 Integrals: definite (see Definite integrals) improper, 108 indefinite (see Indefinite integrals) line, 124 multiple, 125 surface, 125 Integration, 64 (See also Integrals) constant of, 67 general rules, 67–69 Integration by parts, 67 generalized, 69 Intercepts, 22 Interest, 272–275 Intermediate Value Theorem, 233 Interpolation, 227 Hermite, 229 Interpolatory formula (general), 228 Interquartile range, 211 Interval of convergence, 138 Inverse: hyperbolic functions, 59–61 Laplace transforms, 180 trigonometric functions, 49–51 Iteration methods, 240 for general linear systems, 240 for Poisson equation, 240 Jacobi method, 240 Jacobi’s elliptic functions, 199 Jacobian, 128 Ker and Kei functions, 158–159 Kurtosis, 212 Lagrange: form of remainder, 138 interpolation, 227 Laguerre polynomials, 172 generating function for, 173 recurrence formula, 192 Laguerre’s associated differential equation, 170 Laguerre’s differential equation, 172 Landen’s transformation, 199 Laplace transform, 180–192 complex inversion formula for, 180 definition of, 180 inverse, 180 tables of, 181–192 Laplacian, 123, 128 Least-squares: curve, 215 line, 214 Legendre functions, 164–168 of the second kind, 166 Legendre polynomial, 164–165, 232 generating function for, 164 recurrence formula for, 166 tables of values for, 269 Legendre’s associated differential equation, 168 Legendre’s differential equation, 118, 164 Leibniz’s rule, 64 Lemniscate, 28 Limacon of Pascal, 32 Line, 22, 35 of best fit, 214 regression, 214 Line integral, 124
287
Logarithmic functions, 53–55 (See also Logarithms) series for, 139 table of values, 245–246, 252–253 Logarithms, 53–55 of complex numbers, 55 Griggsian, 53 Maclaurin series, 138 Mean, 208 continuous random variable, 224 deviation (M.D.), 211 discrete random variable, 223 geometric, 209 grand, 209 harmonic, 209 population, 212 weighted. 209 Mean value theorem, for definite integrals, 108 generalized, 109 Median, 208 Midpoint rule, 231, 235 Midrange, 210 Milne’s method, 236 Minkowski’s inequality, 206 for integrals, 206 Mode, 209 Modified Bessel functions, 155–157 generating function for, 157 graphs of, 159 recurrence formulas for, 157 Modulus of a complex number, 11 Moment, rth, 212 Momental skewness, 212 Moments of inertia, 41 Monoticity Rule (Probability), 218 Mutinomial coefficients, 9 Multiple, integrals, 125 Napier’s rules, 52 Natural logarithms and antilogarithms, 53 tables of, 252–253 Neumann’s function, 153 Newton’s: backward-difference formula, 228 forward-difference formula, 228 interpolation, 227 method, 233 Nonhomogeneous differential equation, linear second order, 117 Nonlinear equations, solution of, 233 Normal curve, 276–277 distribution, 226 Normal equations for least-squares line, 214 Null function, 189 Numbers: Bernoulli, 142 Euler, 142 Numerical methods for partial differential equations, 237–239 Oblate spheroidal coordinates, 131 Orthogonal curvilinear coordinates, 127–128 formulas involving, 128 Orthogonality: Chebyshev’s polynomials, 176 Laguerre polynomials, 172 Legendre polynomials, 165 Ovals of Cassini, 32
INDEX
288
Parabola, 25 segment of, 18 Parabolic cylindrical coordinates, 129 Paraboloid, 40 Paraboloidal coordinates, 130 Parallelepiped, 19 Parallelogram, 7 Parameter, 208 Parseval’s identity for: Fourier series, 144 Fourier transform, 194 Partial: derivatives, 65 differential equations, numerical methods, 237 Pascal’s triangle, 8 Percentile, kth, 211 Periods of elliptic functions, 200 Plane analytic geometry, formulas from, 22–27 Plane, complex, 10 Poisson: distribution, 226 equation, 237 summation formula, 137 Polar: coordinates, 24 form of a complex number, 11 Polygon, regular, 17 Polynomial function (least-squares), 214 Polynomials: Chebyshev’s, 175 Laguerre, 171 Legendre, 164 Population, 208 mean 210 standard deviation, 212 variance, 212 Power function (least-squares), 214 Power series, 138–141 reversion of 141 Powers, sums of, 134 Present value, of an amount, 273 of an annuity, 275 Probability, 217 distribution, 223 function, 218 tables, 276 Products, infinite, 207 special, 5 Pulse function, 192 Pyramid, volume of, 20 Quadrants, 43 Quadratic convergence, 233 Quadratic equation, solution of, 103 Quadrature, 231–232 Quartic equation, solution of, 13 Quartile coefficient of skewness, 212 Quartiles [QL, M, QU], 211 Radians, 4, 44 table of conversion to degrees, 250 Random numbers table, 282 Random variable, 223–226 standardized, 226 Range, sample, 210 Real part of a complex number, 10 Reciprocals of powers, sums of, 135 Rectangle, 13 Rectangular coordinate system, 120 Rectangular coordinates, 24 transformation to polar coordinates, 24 Rectangular formula, 109 Rectified sine wave function, 191
Recurrence or recursion formulas: Bessel functions, 154 Chebyshev’s polynomials, 175 gamma function, 149 Hermite polynomials, 169 Laguerre polynomials, 171 Legendre polynomials, 165 Regression line, 214 Regular polygon, 17 Remainder: Cauchy’s form, 13 Lagrange form, 138 Remainder formula: Gauss-Legendre interpolation, 232 Hermite interpolation, 230 Lagrange interpolation, 227 Reversion of power series, 141 Richardson method, 240 Riemann zeta function, 204 Right circular cone, 20 Rochigue’s formula: Laguerre polynomials, 171 Legendre’s polynomials, 164 Root mean square (R.M.S.), 211 Roots of complex numbers, 11 Rose, 29 Rotation, 24, 37 Runge-Kutta method, 236 Sample, 208 covariance, 213 Saw tooth wave function, 191 Scalar, 119 multiplication of vectors, 119 Scalar or dot product, 120 Scale factors, 127 Scatterplot, 212 Schwarz (Cauchy-Schwarz) inequality, 205 for integrals, 206 Secant method, 233 Second-order differential equation, 117 Second-order divided-difference formula, 228 Sector of a circle, 17 Segment: of circle, 18 of parabola, 18 Semi-interquartile range, 211 Separation of variables, 116 Series, arithmetic, 134 arithmetic-geometric, 134 binomial, 188 of constants, 134 Fourier, 144–148 geometric, 134 power, 138 of sums of powers, 134 Taylor, 138–141 Simpson’s formula, 109, 231 Sine, 43 graph of, 46 table of values, 247 Sine integral, 88 table of values, 264 Sines, law of, 51 Skewness, 212 Solid analytic geometry, 34–40 Solutions of algebraic equations, 13–14 SOR (successive-overrelaxation) method, 240 Sphere, equations of, 38 surface area, 19 volume, 21 Spherical coordinates, 38, 129 Spherical triangle, 51
INDEX
Spiral of Archimedes, 33 Square wave function, 191 Squares error, 215 Standard deviation, 210 continuous random variable, 225 discrete random variable, 224 population, 212 sample, 210 Standardized random variable, 215 Statistics, 208–216 tables, 276–281 Step function, 192 Stirling’s formula, 150 Stochastic process, 219 Stokes’ theorem, 126 Student’s t distribution, 226 table of, 298 Successive-overrelaxation (SOR) method, 240 Summation formula: Euler-Maclaurin, 137 Poisson, 137 Surface integrals, 125 Tangent function, 43 graph of, 46 table of values, 249 Tangents, law of, 51, 52 Taylor series, 138–141 two variables, 141 Three-point interpolatory formula, 228 Toroidal coordinates, 132 Torus, surface area, volume, 18 Total probability, Law of, 220 Tractrix, 31 Transformation: Jacobian of, 128 of coordinates, 24, 36–37, 128 of integrals, 70, 128 Translation of coordinates: in a plane, 24 in space, 36 Trapezoid, area, perimeter, 16 Trapezoidal rule (formula), 109, 231, 235 Tree diagrams, Probability, 219
289
Triangle inequality, 205 Triangular wave function, 191 Trigonometric functions, 43–52 definition of, 43 graphs of, 46 inverse, 49–50 series for, 139 tables of, 247–249 Triple integrals, 125 Trochoid, 30 Two-point formula, 228 Two-point interpolatory formula, 228 Unit function, Heaviside’s, 192 Unit normal to the surface, 125 Unit vector, 120 Variance, 210 continuous random variable, 225 discrete random variable, 224 population, 210 sample, 210 Vector analysis, 119–133 Vector or cross-product, 121 Vectors, 119 derivatives of, 122 integrals involving, 124 unit, 119 Volume integrals, 125 Wallis’ product, 207 Wave equation, 238 Weber’s function, 153 Weighted mean, 209 Witch of Agnesi, 31 x-intercept, 22 y-intercept, 22 Zero vector, 119 Zeros of Bessel functions, 267 Zeta function of Riemann, 204