Qinghai Zhang 1 Summary of Math 1321-004 2013-MAR-18 Fundamentals Definition 7. A function f from X to Y, written as f : X → Y or X 7→ Y, is a subset of the Cartesian product X × Y satisfying that ∀x ∈ X , there is exactly one y ∈ Y s.t. (x, y) ∈ X × Y. X and Y are the domain and range of f , respectively. Nomenclature • R, N: the sets of real and natural numbers. • calligraphic uppercase letters S, U, P: sets, • sans serif uppercase letters U, E: statements, Definition 8. A function f : X → Y is said to be injective or one-to-one iff • uppercase letters A, B, P : points, • lowercase boldfaced letters a, v: vectors, ∀x1 ∈ X , ∀x2 ∈ X , x1 6= x2 ⇒ f (x1 ) 6= f (x2 ). • lowercase letters x, y, c, d: scalars, or set elements, It is surjective or onto iff • lowercase letters m, n: natural numbers Definition 1. A set S is a collection of distinct objects x’s, often denoted with the following notation S = {x | the conditions that x satisfies. }. (9) ∀y ∈ Y, ∃x ∈ X , s.t. y = f (x). (10) It is bijective iff it is both injective and surjective. (1) Definition 2. A universal statement is a logic statement Definition 9. A set S is countable iff there exists an injective function f : S → N that maps S to N. of the form U = (∀x ∈ S, A(x) holds). An existential statement has the form E = (∃x ∈ S, s.t. A(x) holds). (2) Definition 10. A 3D coordinate system is a bijective function whose domain and range are R3 and P, where R3 = R × R × R and P is the set of all spatial locations. (3) Definition 11. The distance between two points A = (a1 , a2 , a3 ) and B = (b1 , b2 , b3 ) is A statement of implication/conditional has the form A ⇒ B. (4) |AB| = p (a1 − b1 )2 + (a2 − b2 )2 + (a3 − b3 )2 . (11) Axiom 3 (First-order negation of logical statements). The negations of the statements in Definition 2 are ¬U = (∃x ∈ S, s.t. A(x) does not hold), ¬E = (∀x ∈ S, A(x) does not hold). Definition 12. Given two points A = (a1 , a2 , a3 ) and B = (b1 , b2 , b3 ), the vector v = B − A is the displace(5) ment of B from A. The length of vector v equals the (6) distance between A and B: Definition 4. S is a subset of U, written as S ⊆ U, iff x ∈ S ⇒ x ∈ U. |v| = |AB|. In particular, a = ha1 , a2 , a3 i can be regarded as a position vector that starts at the origin O = (0, 0, 0) and ends at A. Definition 5. The Cartesian product X × Y between two sets X and Y is the set of all possible ordered pairs with first element from X and second element from Y: X × Y = {(x, y) | x ∈ X , y ∈ Y}. (12) (7) a = ha1 , a2 , a3 i ⇒ |a| = q a21 + a22 + a23 . (13) Axiom 6 (Fundamental principle of counting). A task consists of a sequence of k steps. Let ni denote the num- Axiom 13 (Vector addition and scaling). If c ∈ R, ber of different choices for the i-th step, the total number u = hxu , yu , zu i, v = hxv , yv , zv i, then of distinct ways to complete the task is then u + v = hxu + xv , yu + yv , zu + zv i , (14a) k Y cv = hcvx , cvy , cvz i . (14b) ni = n1 n2 · · · nk . (8) i=1 1 Qinghai Zhang Summary of Math 1321-004 2013-MAR-18 Axiom 14 (Vector algebra). If c ∈ R and u, v, w ∈ Rn , Definition 24. A set is an open set if it contains none then of its boundary points. u + v = v + u, u + (v + w) = (v + u) + w, (15a) Definition 25. A set is a closed set if it contains all of (15b) its boundary points. Definition 26. A point set U ⊆ Rn is bounded iff n u + (−u) = 0 (15d) U ⊆ B(P0 , r) for some P0 ∈ R and r > 0. c(u + v) = cv + cu, (15e) Definition 27 (Limit of a scalar function with multiple variables). The limit of a function f : B0 (P0 , r) → R ex(c + d)u = cu + du (15f) ists as P approaches P0 , written as limP →P0 f (P ) = L, (cd)u = c(du), (15g) iff 1u = u. (15h) ∀ > 0, ∀ paths P → P0 , ∃δ > 0, s.t. (22) Definition 15. A scalar function is a function whose ∀P ∈ B(P0 , δ), |f (P ) − L| < . range is a subset of R. Formula 28. The limit of f : B0 (P0 , r) → R does not Definition 16. A vector function is a function whose exist at P0 if there exists two different paths P → P0 and range is a subset of Rn with n > 1. P P0 s.t. Definition 17. A vector field is a vector function (23) lim f (P ) = L1 6= L2 = lim f (P ) . m m P P0 P →P0 F:R →R . u+0=u (15c) Definition 18 (Limit of a scalar function with one vari- Theorem 29 (The squeeze theorem). If ∃r > 0 s.t. able). Consider a function f : I → R with I(a, r) = ∀P ∈ B0 (P0 , r), f (P ) ≤ g(P ) ≤ h(P ), then (a − r, a) ∪ (a, a + r). The limit of f (x) exists as x aplim f (P ) = lim h(P ) = L ⇒ lim g(P ) = L. proaches a, written as limx→a f (x) = L, iff P →P0 ∀ > 0, ∃δ, s.t. ∀x ∈ I(a, δ), |f (x) − L| < . P →P0 P →P0 (16) (24) Definition 19 (Limit of a scalar function with one vari- Definition 30. f : R → R is continuous at a iff able at infinity). Consider a function f : (a, ∞) → R. lim f (x) = f (a). (25) x→a limx→∞ f (x) = L iff Definition 31. f : Rn → R is continuous at Q iff ∀ > 0, ∃M ∈ (a, ∞), s.t. ∀x > M, |f (x) − L| < . (17) lim f (P ) = f (Q). (26) P →Q Definition 20 (Limit of a vector function with one variable). If r : R → Rn , r(t) = hr1 (t), r2 (t), . . . , rn (t)i , f is continuous on a point set U if (26) holds ∀Q ∈ U. then Definition 32. The derivative of a function f : R → R D E lim r(t) = lim r1 (t), lim r2 (t), · · · , lim rn (t) . (18) at a is the limit t→a t→a t→a t→a f (a + h) − f (a) f 0 (a) = lim . (27) Definition 21. The open ball centered at P0 ∈ Rn with h→0 h radius r > 0 is the point set If the limit exists, f is differentiable at a. B(P0 , r) = P |P − P0 | < r . (19) Definition 33. A function f (x) is C k or k times conIt is an open interval in 1D and an open disk in 2D. tinuously differentiable if f (k) (x) exists and is itself conThe open ball without the center is denoted by tinuous. B0 (P0 , r) = B(P0 , r) \ {P0 }. (20) Theorem 34. A polynomial is continuous everywhere. Definition 22. f : Rn → R has a local maximum at Definition 35. A sequence is simply a countable set P0 ∈ Rn iff {an | n ∈ N+ } or a function defined on N+ . ∃r > 0, s.t. ∀P ∈ B(P0 , r), f (P ) ≤ f (P0 ). Changing ≤ to ≥ in (21) yields a local minimum. An extremum is either a maximum or minimum. (21) Definition 36 (Limit of a sequence). A sequence {an } has the limit L, written as limn→∞ an = L, or an → L as n → ∞, iff ∀ > 0, ∃N, s.t. ∀n > N, |an − L| < . Definition 23. P0 is a boundary point of a point set U iff ∀r > 0, ∃P ∈ B(P0 , r) s.t. P 6∈ U. (28) If such a limit L exists, we say that {an } converges to L. 2 Qinghai Zhang 2 Summary of Math 1321-004 2013-MAR-18 Infinite sequences and Series Definition 37. A series is the sum of all terms in a sequence {an }. P∞ n−1 Theorem 38. The geometric series with n=1 ar a 6= 0 is convergentPif |r| < 1 and divergent otherwise. ∞ a . In the former case, n=1 arn−1 = 1−r P∞ 1 Theorem 39. The p-series n=1 np is convergent if p > 1 and divergent if p ≤ 1. P∞ Theorem 40. n=1 an converges ⇒ limn→∞ an = 0. Theorem 50 (Remainder estimate for integral test). Let f be a function that satisfies the conditions in Theorem 42, then ∀n > M , Z ∞ Z ∞ f (x)dx. (31) f (x)dx ≤ Rn ≤ n n+1 Theorem 51 (Alternating series estimation). A convergent alternating series defined in Theorem 45 satisfies |Rn | ≤ bn+1 . Definition 52. A power series centered at a is a series of the form ∞ Theorem 41 (Test for divergence). X P∞If limn→∞ an does p(x) = cn (x − a)n , (32) not exist or limn→∞ an 6= 0, then n=1 an diverges. n=0 Theorem 42 (Integral test). Let f : R → R be a con- where cn ’s are the coefficients. The interval of convertinuous bounded function satisfying f (n) = an . gence is the set of x values for which the series converges: 0 If ∃MP∈ N+ , s.t. ∀x ∈ [M, +∞), f (x) > 0 and f (x) < 0, R∞ ∞ Ic (p) = {x | p(x) converges}. (33) then n=1 an converges iff M f (x)dx is convergent. Theorem 53 (Interval of convergence of power series). Theorem 43P(Comparison test). Consider two series P∞ There are only three possibilities for Ic (p) of the power ∞ + n=1 an and n=1 bn satisfying ∀n ∈ N , 0 < an ≤ bn . series (32): (i) Ic = {a}, (ii) Ic = (−∞, +∞), P∞ P∞ (iii) Ic = (a − R, a + R) or [a − R, a + R) or (a − R, a + R] • If n=1 bn converges, then n=1 an converges. or [a − R, a + R]. P∞ P∞ • If n=1 an diverges, then n=1 bn diverges. R is called the radius of convergence. R = 0, +∞ for cases (i) & (ii), respectively. Theorem 44P(Limit comparison test). If two series P ∞ ∞ + Theorem 54 (Term-by-term differentiation and integraP∞ n=1 an and n=1 bn satisfy ∀n ∈ N , an > 0, bn > 0, tion). If f (x) = c (x − a)n has radius of convern and n=0 an = c ∈ (0, ∞), (29) gence R > 0, then f (x) is differentiable on (a − R, a + R). lim n→∞ bn ∞ X then either both series converge or both diverge. f 0 (x) = ncn (x − a)n−1 , (34a) n=0 Theorem 45 (AlternatingP series test). An alternating ∞ series is a series of the form n=1 (−1)n−1 bn with bn > 0 + ∀n ∈ N . If it satisfies (i) ∀n ≥ 1, bn+1 ≤ bn , and (ii) P∞ limn→∞ bn = 0, then n=1 (−1)n−1 bn converges. P∞ Definition 46. A series P∞ n=1 an is called absolutely convergent if the series n=1 |an | converges. P∞ Theorem 47P(Absolute convergence). If n=1 |an | con∞ verges, then n=1 an converges. P∞ Theorem 48 (Ratio test). Consider a series Pn=1 an ∞ an+1 satisfying limn→∞ | an | = L. If L < 1, then n=1 an P∞ converges; if L > 1, then n=1 an diverges; if L = 1, no conclusion can be drawn. P∞ Definition 49. The nth remainder of the series i=1 ai is defined as Rn = ∞ X i=1 ai − n X i=1 ai = ∞ X ai . Z f (x)dx = C + ∞ X cn n=0 (x − a)n+1 . n+1 (34b) Furthermore, the radii of convergence of both f 0 (x) and R f (x)dx are R. Definition 55. If f (n) (x) exists for a function f : R → R at x = a, then Tn (x) = n X f (k) (a) k=0 k! (x − a)k (35) is called the nth Taylor polynomial for f (x) at a. In particular, the linear approximation for f (x) at a is T1 (x) = f (a) + f 0 (a)(x − a). (36) Definition 56. The Taylor series (or Taylor expansion) for f (x) at a is (30) lim Tn (x) = n→∞ i=n+1 3 ∞ X f (k) (a) k=0 k! (x − a)k . (37) Qinghai Zhang Summary of Math 1321-004 2013-MAR-18 Definition 57. The remainder of the nth Taylor poly- Definition 62. ∀k, n ∈ N, the binomial coefficients are nomial in approximating f (x) is ( k! if k ≥ n, k Rn (x) = f (x) − Tn (x). (38) (42) = n!(k−n)! n 0 if k < n. Theorem 58. Let Tn be the nth Taylor polynomial for f (x) at a. ∀k ∈ R, ∀n ∈ N, the binomial coefficients are lim Rn (x) = 0 ⇔ f (x) = lim Tn (x). (39) ( Qn−1 (k−i) n→∞ n→∞ i=0 k if n > 0, n! = (43) (m) Lemma 59. ∀m = 0, 1, 2, . . . , n, Rn (a) = 0. n 1 if n = 0. Theorem 60 (Taylor’s theorem with Lagrangian form). Consider a function f : R → R. If f (n+1) (x) exists on Theorem 63 (Binomial series). ∀k ∈ R, |x| < 1, the interval I = (a − d, a + d), and f (n) (x) is continuous ∞ ∞ Qn−1 X X k n i=0 (k − i) n on [a − d, a + d], then ∀x ∈ I, ∃y ∈ I s.t. (1 + x)k = x (44) x =1+ n! n (n+1) n=0 n=1 f (y) ∞ (x − a)n+1 . (40) Rn (x) = X k(k − 1) · · · (k − n + 1) n (n + 1)! x =1+ n! Theorem 61 (Taylor’s inequality). If ∃M < ∞, s.t. n=1 ∀x ∈ [a−d, a+d], f (n+1) (x) ≤ M , then ∀x ∈ (a−d, a+d), |Rn (x)| ≤ M |x − a|n+1 . (n + 1)! (41) If ∀n ∈ N, (41) holds, then limn→∞ Tn = f (x). 4 Qinghai Zhang 3 Summary of Math 1321-004 2013-MAR-18 Vectors and Curves Definition 73. The standard basis vectors in R3 are e1 = i = h1, 0, 0i , is a unit vector in the same direction e2 = j = h0, 1, 0i , (54) e3 = k = h0, 0, 1i . Definition 64. v is a unit vector iff |v| = 1. Formula 65. of v. v |v| Definition 66. A line is a set of points uniquely determined by a point P0 and a direction vector v: {P | P (t) = P0 + tv, t ∈ (−∞, +∞)}. (45) a · b = |a||b| cos θ. (48) Definition 74 (Geometric definition of cross product). The cross product of two vectors a, b ∈ R3 is a × b = (|a||b| sin θ)n, (55) Definition 67 (Dot product: algebraic definition). The where θ is the angle between a and b, and n is the unit dot product of two vectors a, b ∈ R3 is a real number: vector determined by the right-hand rule from a and b. a · b = a 1 b1 + a 2 b2 + a 3 b3 . (46) Definition 75 (Algebraic definition of cross product). Definition 68. The angle θ between two nonzero vec tors a, b ∈ R3 satisfies i j k a × b = det a1 a2 a3 (56a) a·b b1 b2 b3 cos θ = , θ ∈ [0, π]. (47) |a||b| a2 a3 a1 a3 a1 a2 Theorem 69. The algebraic definition of the dot prod- = det b2 b3 i − det b1 b3 j + det b1 b2 k uct is equivalent to its geometric definition: (56b) = (a2 b3 − a3 b2 )i − (a1 b3 − a3 b1 )j + (a1 b2 − a2 b1 )k (56c) Theorem 70 (Algebra of dot product). If u, v, w ∈ R3 , Theorem 76. The algebraic and geometric definitions and c ∈ R, then of cross product are equivalent. u · u = |u|2 (49a) u·v =v·u (49b) Theorem 77 (Algebra of cross product). If c ∈ R and 3 u · (v + w) = u · v + u · w (49c) u, v, w ∈ R , then (cu) · v = c(u · v) = u · (cv) 0·u=0 (49d) u × v = −v × u (49e) Definition 71. The scalar projection of b onto a is compa b = a·b , |a| (57b) u × (v + w) = u × v + u × w (57c) (u + v) × w = u × w + v × w (57d) (50) Definition 78. Two nonzero vectors a, b are perpendicular or orthogonal, written as a ⊥ b, iff a · b = 0. and the vector projection of b onto a is proja b = ca = (compa b) (57a) (cu) × v = c(u × v) = u × (cv) a . |a| n (51) Definition 79. Two nonzero vectors a, b ∈ R are parallel, written as a k b, iff ∃c 6= 0, s.t. a = cb. Definition 72. A plane is a set of points uniquely de- Theorem 80. a, b ∈ R3 are parallel iff a × b = 0. termined by a point P0 and a normal vector n = ha, b, ci: {P | n · (P − P0 ) = 0}. Definition 81 (scalar triple product). For a, b, c ∈ R3 , (52) a1 a · (b × c) = (a × b) · c = det b1 c1 Equivalently, the scalar equation of a plane is ax + by + cz + d = 0. (53) 5 a2 b2 c2 a3 b3 c3 . (58) Qinghai Zhang Summary of Math 1321-004 2013-MAR-18 Theorem 82. For u, v : R → R3 , c ∈ R, f : R → R, d [u(t) + v(t)] = u0 (t) + v0 (t), dt d [cu(t)] = cu0 (t), dt d [f (t)u(t)] = f 0 (t)u(t) + f (t)u0 (t), dt d [u(t) · v(t)] = u0 (t) · v(t) + u(t) · v0 (t), dt d [u(t) × v(t)] = u0 (t) × v(t) + u(t) × v0 (t), dt d [u(f (t))] = f 0 (t)u0 (f ). dt Definition 90. The curvature of a curve r(t) at the point P (t) = O + r(t) is (59a) dT . (64) κ(t) = ds (59b) Formula 91. (59c) κ(t) = (59d) (59e) κ(t) = (59f) |r0 (t) × r00 (t)| |r0 (t)|3 (66) Corollary 93. The curvature of a 2D curve y = f (x) is κ(x) = Definition 84. A surface is (the image of) a vector function R2 7→ R3 . |f 00 (x)| [1 + (f 0 (x))2 ] 3/2 . (67) Definition 94. The principal unit normal vector is N(t) = Definition 85. The tangent vector to a curve r(t) = hr1 (t), r2 (t), r3 (t)i at a point P (t) = O + r(t) is T0 (t) , |T0 (t)| (68) and the binormal vector is r(t + ∆t) − r(t) = hr10 (t), r20 (t), r30 (t)i , ∆t→0 ∆t (60) r0 (t) = lim B(t) = T(t) × N(t). (69) The normal plane of the curve at P = O + r(t) is the plane determined by N(t) and B(t). The osculating plane is that by T(t) and N(t). the corresponding unit tangent vector is r0 (t) . |r0 (t)| (65) Theorem 92. The curvature of a curve r(t) at P (t) is Definition 83. A curve is (the image of) a vector function R 7→ R3 , or r(t) : R → R3 . The independent variable t is its parametrization. T(t) = |T0 (t)| . |r0 (t)| (61) Definition 95. Let t represent time and r(t) the trajectory of a moving particle. Then r0 (t) = v is called the Definition 86. The tangent line to r : R → R3 at velocity of the particle, |r0 (t)| = |v| = v the speed of the P (t0 ) = O + r(t0 ) is the line determined by P (t0 ) and T: particle, r00 (t) = a the acceleration of the particle. {P | P = P (t0 ) + tT, t ∈ R}. (62) Theorem 96. The acceleration of a particle following the curve r(t) is a vector a(t) consists of two parts: Theorem 87. If |r(t)| = c where c is a constant, then r(t) · r0 (t) = 0. Consequently r(t) · T(t) = 0. a(t) = aT T + aN N, (70) Definition 88. The arc length of a curve r : R → R3 starting from P (a) = O + r(a) is a function s : R → R, Z s(t) = where aT is caused by the change of the speed, and aN is caused by the change of the velocity direction: t |r0 (u)|du. r0 · r00 , |r0 | |r0 × r00 | aN (t) = κv 2 = . |r0 | (63) aT (t) = v 0 = a Formula 89. ds = |r0 (t)|. dt 6 (71a) (71b) Qinghai Zhang 4 Summary of Math 1321-004 2013-MAR-18 Partial Derivatives Definition 97. The partial derivative of f : Rn → R Definition 102. The total differential of f : Rn → R is with respect to the ith dimension at P0 is defined in terms of the differential s dxi ’s: n f (P0 + hei ) − f (P0 ) ∂f (P0 ) ∂f X ∂f = lim = , (72) df = dxi (79) ∂xi ∂xi P0 h→0 h ∂xi i=1 where h ∈ R and {ei | i = 1, 2, · · · , n} is the set of stanDefinition 103. The increment of a function z = dard basis vectors of the Euclidean n-space, i.e. the ith f (x, y) at a point P0 = (x0 , y0 ) is component of ei is 1 and all other components are 0. ∆z = f (x, y) − f (x0 , y0 ). Theorem 98 (Clairaut’s). If f : Rn → R has continuous second partial derivatives at P , then ∀i, j = 1, 2, · · · , n, 2 2 ∂ f (P ) ∂ f (P ) = . ∂xi ∂xj ∂xj ∂xi Definition 104. f : B(P0 , r) → R is differentiable at P0 if there exists a linear function L(r) = c · r where c ∈ Rn (73) such that for all paths P → P , 0 Definition 99. Big O notation describes the limiting behavior of a function in terms of another function. Given f, g : R → R, f (h) = L 6= 0 g(h) f (h) f (h) = o(g(h)) ⇔ lim =0 h→0 g(h) f (h) = O(g(h)) ⇔ lim h→0 (80) lim P →P0 f (P ) − f (P0 ) − L(P − P0 ) = 0. |P − P0 | (81) Theorem 105. f : Rn → R is differentiable at P iff ∂f ∀i ∈ 1, 2, · · · , n, ∂x exists and is continuous at P . i (74a) Definition 106. A C 1 function f : R2 → R admits a (74b) pair of tangent vectors along x- and y- directions for the surface z = f (x, y): Definition 100. The linear approximation of f (x, y) at ∂z ∂z P0 = (x0 , y0 ) is u = 1, 0, , v = 0, 1, . (82) ∂x ∂y ∂f ∂f T1 (x, y) = f (x0 , y0 ) + (x − x0 ) + (y − y0 ), Definition 107. The tangent plane of the surface at P0 ∂x P0 ∂y P0 is the plane that contains P0 and the tangent vectors (75) u(P0 ) and v(P0 ). which satisfies Lemma 108. If f : R2 → R is C 1 at z0 = f (x0 , y0 ), the scalar equation of the tangent plane of the surface σ(x, y) = hx, y, f (x, y)i at (x0 , y0 , z0 ) is where h = max (|x − x0 |, |y − y0 |). ∂z ∂z Definition 101. The quadratic approximation of f (x, y) z − z0 = (x − x ) + (y − y0 ). (83) 0 ∂x P0 ∂y P0 at P0 = (x0 , y0 ) is n ∂f ∂f T2 (x, y) =f (x0 , y0 ) + (x − x ) + (y − y0 ) Theorem 109 (The chain rule). If f : R → R is dif0 ∂x P0 ∂y P0 ferentiable and each variable xi is a differentiable func 2 2 2 2 tion of m variables t1 , t2 , · · · , tm , then f is a function of ∂ f (y − y0 ) ∂ f (x − x0 ) + + t 1 , t2 , · · · , tm and ∀i = 1, 2, · · · , m, ∂x2 P0 2 ∂y 2 P0 2 n X ∂ 2 f ∂f ∂f ∂xj + (x − x0 )(y − y0 ), (77) = . (84) ∂x∂y P0 ∂ti ∂x j ∂ti j=1 f (x, y) = T1 (x, y) + O(h2 ) (76) which satisfies f (x, y) = T2 (x, y) + O(h3 ) Definition 110. A level curve of F : R2 → R is the curve with equation F (x, y) = c where c is a constant. A level surface of F : R3 → R is the surface with equation F (x, y, z) = c where c is a constant. (78) where h = max (|x − x0 |, |y − y0 |). 7 Qinghai Zhang Summary of Math 1321-004 Theorem 111 (Implicit function theorem: 2D). Con∂F 2 sider F : R → R at P0 = (x0 , y0 ). If ∂y 6= 0 and F 2013-MAR-18 Theorem 119 (Fermat’s). Every local extremum of f : Rn → R is a critical point. P0 is C 1 on the open disk B(P0 , r) for some r > 0, then the Formula 120 (Second derivatives test for critical level curve F (x, y) = c defines a C 1 function y = y(x) on points). For a critical point P0 = (a, b) of f : R2 → R, 2 2 ∂2f the open interval B(x0 , r), and if ∂∂xf2 , ∂∂yf2 , and ∂x∂y are all continuous on B(P0 , r) for some r > 0, then the discriminant ∂F dy ∂x = − ∂F . (85) 2 dx ∂ f ∂2f ∂y ∂x2 ∂x∂y (92) D = det 2 2 ∂ f ∂ f Theorem 112 (Implicit function theorem: 3D). Con ∂y∂x 2 ∂y P0 sider F : R3 → R at P0 = (x0 , y0 , z0 ). If ∂F 6= 0 ∂z P0 and F is C 1 on the open ball B(P0 , r) for some r > 0, might determine the then the level surface F (x, y, z) = c defines a C 1 function z = z(x, y) on the open disk B((x0 , y0 ), r), and D < 0, 2 D > 0, ∂∂xf2 > 0 ∂F ∂F ∂z ∂z ∂y P 0 ∂x 2 , (86) = − ∂F = − ∂F . D > 0, ∂∂xf2 < 0 ∂x ∂y ∂z ∂z P0 type of the critical point: ⇒ f (P0 ) is not a local extremum, ⇒ f (P0 ) is a local minimum, ⇒ f (P0 ) is a local maximum. (93) Definition 113. The gradient vector of a function f : n n R → R is a vector in R : Theorem 121. If f : Rn → R is continuous on a closed, X n bounded set D ⊆ Rn , then f attains an absolute maxi∂f ∂f ∂f ∂f ∇f = , ,··· , = ei , (87) mum value and an absolute minimum value on D. ∂x1 ∂x2 ∂xn ∂xi i=1 Theorem 122. Consider determining extrema of a C 1 function f : Rn → R under the constraint F (P ) = c where F : Rn → R, F ∈ C 1 , c is a constant. If f attains an extremum at P0 and ∂F∂x(Pn0 ) 6= 0, then ∃λ ∈ R s.t. Definition 114. The directional derivative of a function f : Rn → R at P0 ∈ Rn in the direction of a vector u ∈ Rn is the scalar Du f (P0 ) = lim h→0 f (P0 + hu) − f (P0 ) . h (88) ∇f |P0 = λ ∇F |P0 . (94) Theorem 115. If f : Rn → R is differentiable at P , Definition 123. Consider determining extrema of a C 1 then function f (x, y, z) under the constraint F (x, y, z) = c Du f (P ) = u · ∇f (P ). (89) where F ∈ C 1 and c is a constant. By constructing the Lagrangian function Theorem 116. If f : Rn → R is differentiable at P , then Λ(x, y, z, λ) = f (x, y, z) + λ c − F (x, y, z) , (95) max Du f (P ) = |∇f (P )|. (90) n u∈R , |u|=1 the method of Lagrange multipliers first find the point Lemma 117. Let z = z(x, y) be the implicit function set U = {(x, y, z) | ∇Λ(x, y, z, λ) = 0} by solving defined by the level surface F (x, y, z) = c. The unit normal vector of the tangent plane to the level surface ∇f = λ∇F, (96) F (x, y, z) = c is F (x, y, z) = c. ∇F n= . (91) and then evaluate f over U to obtain |∇F | Definition 118. A critical point of f : Rn → R is a point P0 ∈ Rn satisfying ∇f |P0 = 0. max f = max f (P ), P ∈U 8 min f = min f (P ). P ∈U (97) Qinghai Zhang 5 Summary of Math 1321-004 2013-MAR-18 Multiple integral Definition 132. The Riemann sum of f : R2 → R over a partition Tm,n is Definition 124. A partition of an interval I = [a, b] is a finite ordered subset Tn ⊆ I of the form Tn (a, b) = {a = x0 < x1 < · · · < xn = b}. (98) Sm,n (f ) = m X n X ∗ f x∗ij , yij ∆Aij , (106) i=1 j=1 The interval Ii = [xi−1 , xi ] is the ith subinterval of the partition. The norm of the partition is the length of the longest subinterval, ∗ ∈ Rij where ∆Aij = (xi − xi−1 )(yj − yj−1 ) and x∗ij , yij is the sample point of the (i, j)th subrectangle. hn = h(Tn ) = max(xi − xi−1 ), i = 1, 2, . . . , n. (99) Definition 133. If f : R2 → R is integrable on R = [a, b] × [c, d], then its limit is called the definite integral Definition 125. The Riemann sum of f : R → R over of f on R: a partition Tn is ZZ f (x, y)dA = lim Sm,n (f ). (107) n X m,n→∞ ∗ R Sn (f ) = f (xi )(xi − xi−1 ), (100) i=1 Formula 134. The signed volume of the solid that lies between a surface z = f (x, y) and a rectangle R inside where x∗i ∈ Ii is a sample point of the ith subinterval. the xy-plane is ZZ Definition 126. f : R → R is integrable (or more preV = f (x, y) dA. (108) cisely Riemann integrable) on [a, b] iff R ∃L ∈ R, s.t. ∀ > 0, ∃δ s.t. Definition 135. The average value of f : R2 → R on (101) R = [a, b] × [c, d] is ZZ 1 Definition 127. If f : R → R is integrable on [a, b], f (x, y) dA. (109) hf iR = (b − a)(d − c) then its limit is called the definite integral of f on [a, b]: R ∀Tn (a, b) with h(Tn ) < δ, |Sn (f ) − L| < . Theorem 136. Consider r ∈ R, R = [a, b] × [c, d], and (102) continuous f, g : R → R. n→∞ a ZZ ZZ ZZ Theorem 128. A continuous function f is integrable (f + g) dA = f dA + g dA, (110a) R R over [a, b]. ZZ ZR Z rf dA = r f dA, (110b) Theorem 129. A monotonic function f is integrable R R Z Z ZZ over [a, b]. (f ≥ g) ⇒ f dA ≥ g dA , (110c) R R Definition 130. A closed rectangle R ⊂ R2 is the CarteZZ sian product of two closed intervals [a, b] and [c, d]: rdA = r(b − a)(d − c). (110d) R R = (x, y) ∈ R2 | x ∈ [a, b], y ∈ [c, d] . (103) Theorem 137 (Fubini’s). If f : R2 → R is continuous Formula 131. A 2D partition of a closed rectangle can on R = [a, b] × [c, d], then Z b f (x)dx = lim Sn (f ). be obtained by the Cartesian product of two 1D partitions: ZZ R Tm,n (R) = Tm (a, b) × Tn (c, d). Z b Z f dA = (104) d Z d Z f dy dx = a c b f dx dy. (111) c a Corollary 138. If f (x, y) = g(x)h(y) is continuous on Any subrectangle is a smaller rectangle Ri,j = Ii × Ij . R = [a, b] × [c, d], then ! Z ! The norm of the 2D partition is ZZ Z b d f (x, y) dA = g(x) dx h(y) dy . (112) h(Tm,n ) = max h(Tm ), h(Tn ) . (105) R a c 9 Qinghai Zhang Summary of Math 1321-004 2013-MAR-18 Formula 139 (Midpoint rule for a double integral). If f : R2 → R is C 3 over a rectangle R, then Theorem 146 (Change of variables in a double integral). Consider a C 1 , bijective function F : S → D that maps a region S ⊂ R2 in the uv-plane to another region ZZ m X n X D ⊂ R2 in the xy-plane. If both S and D are regular, f (x, y) dA = f (x̄i , ȳj ) ∆Aij + O h4m,n , then for a continuous f : D → R, R i=1 j=1 ZZ ZZ (113) f (x, y) dA = f x(u, v), y(u, v) |det(JF )| dudv. where hm,n is the norm of the 2D partition (104), and D S (119) 1 1 ȳj = (yj−1 + yj ). (114) Definition 147. In the polar coordinate system, the pox̄i = (xi−1 + xi ), 2 2 sition of each point P on the Euclidean plane is deterDefinition 140. A planar region D is regular if it can mined by r = |P − O|, its distance from the origin O, be described in either of these two ways and θ, the angle between the two vectors P − O and i. The relations between the polar coordinates (r, θ) and Type I. D = (x, y) | x ∈ [a, b], y ∈ [g1 (x), g2 (x)] ; the Cartesian coordinates (x, y) are x = r cos θ, Type II. D = (x, y) | y ∈ [c, d], x ∈ [h1 (y), h2 (y)] . (120) y = r sin θ. Theorem 141. If f is continuous on a regular region D, then its double integral can be calculated by an iterated Corollary 148. If f (x, y) is continuous on the polar rectangle integral : # ZZ Z "Z D = {(r, θ) | 0 ≤ a ≤ r ≤ b, 0 ≤ α ≤ θ ≤ β ≤ 2π}, b g2 (x) f (x, y) dA = D f (x, y) dy dx, a ZZ Z g1 (x) d "Z f (x, y) dA = D # h2 (y) f (x, y) dx dy. c h1 (y) (121) then its double integral over D can be calculated as ZZ Z βZ b (115b) f (x, y) dA = f (r cos θ, r sin θ) rdr dθ. (115a) D α a (122) Theorem 142. If (i) D1 ∩ D2 is empty or only contains boundary points of D1 and D2 , (ii) f : R2 → R is con- Formula 149. If the density function ρ(x, y) is continuous for a lamina D, then the mass of the lamina is given tinuous on D = D1 ∪ D2 , then by ZZ ZZ ZZ ZZ m= ρ(x, y) dA, (123) f dA. (116) f dA + f dA = D D D2 D1 its moments about the x-axis and y-axis by Definition 143. A function f : R → R is even iff ZZ ZZ f (x) = f (−x); it is odd iff f (x) = −f (−x). Mx = yρ(x, y) dA, My = xρ(x, y) dA, (124) D Formula 144. Let a > 0. ( Ra Z +a 2 0 f (x)dx, f (x)dx = 0, −a its center of mass (x̄, ȳ) by if f (x) is even, (117) if f (x) is odd. x̄ = Definition 145. The Jacobian matrix of a function F : R2 → R2 given by x = x(u, v), y = y(u, v) is ∂x ∂u ∂y ∂u D ∂x ∂v ∂y ∂v My , m ȳ = Mx , m (125) its moment of inertia about the x-axis and y-axis by ZZ ZZ Ix = y 2 ρ(x, y) dA, Iy = x2 ρ(x, y) dA. (126) D D (118) Formula 150. If σ(u, v) = hx(u, v), y(u, v), z(u, v)i is C 1 and is injective on D, then the area of the surface S = {P | P = σ(u, v), (u, v) ∈ D} is and det(JF ) is the Jacobian (or more precisely Jacobian ZZ ∂σ ∂σ determinant). dA. × (127) A(S) = ∂v D ∂u JF = , 10 Qinghai Zhang 6 Summary of Math 1321-004 Vector calculus 2013-MAR-18 Formula 163. If C = {r(t) | t ∈ [a, b]}, then Definition 151. A vector field F : Rm → Rm is C k if each component is k times continuously differentiable. Z Z f (x, y)ds = C b q 2 2 f x(t), y(t) (x0 (t)) + (y 0 (t)) dt. a (136) Definition 152. The vector differential operator ∇ is More generally, for f : Rm → R and r : R → Rm , ∂ ∂ ∂ ∂ ∂ ∂ +j +k = , , . (128) ∇=i Z Z b ∂x ∂y ∂z ∂x ∂y ∂z f (r(t))ds = f r(t) r0 (t)dt. (137) C a Definition 153. The divergence operator of a vector field F : Rm → Rm is Formula 164. If C = {r(t) | t ∈ [a, b]}, then div F = ∇ · F. (129) Z Z b f (x, y)dx = f x(t), y(t) x0 (t)dt, (138a) Definition 154. A vector field F : Rm → Rm is said to C a be incompressible if ∇ · F = 0 holds for all points. Z Z b f x(t), y(t) y 0 (t)dt. (138b) f (x, y)dy = Definition 155. A point P is a source of a vector field a C F if ∇ · F|P > 0; it is a sink of F if ∇ · F|P < 0. m m Definition 156. The curl operator of a vector field Definition 165. Let a vector field F : R → R be continuous on an m-dimensional curve C = {r(t) | t ∈ [a, b]}. F : R3 → R3 is curl F = ∇ × F. (130) The line integral of F along C is Z Z b Z Definition 157. A vector field F : Rm → Rm is said to 0 m F · dr = F r(t) · r (t)dt = F · Tds. (139) be conservative iff ∃f : R → R s.t. F = ∇f . C a C Then f is called a potential function for F. Theorem 158. If a C 1 vector field F : R3 → R3 satisfies Formula 166. If F = hP (x, y, z), Q(x, y, z), R(x, y, z)i, then Z Z ∇ × F = 0, then F is conservative. F · dr = P dx + Qdy + Rdz. (140) Theorem 159. If f : R3 → R has continuous secondC C order partial derivatives, then Theorem 167 (Fundamental theorem of line integrals). ∇ × ∇f = 0. (131) If f : Rm → R is differentiable and ∇f is continuous on an m-dimensional curve C = {r(t) | t ∈ [a, b]}, then Theorem 160. For a C 2 vector field F : R3 → R3 , Z ∇ · (∇ × F) = 0. (132) ∇f · dr = f r(b) − f r(a) . (141) C Definition 161. The line integral of f (x, y) along a curve C (with respect to the arc length) is Corollary 168. Line integrals of conservative vector Z fields are independent of path. n X f (x, y)ds = lim f (x∗i , yi∗ )∆si . (133) n→∞ Definition 169. A curve r(t), t ∈ [a, b] is simple iff r(t) C i=1 is injective on (a, b); it is closed iff r(a) = r(b). It is More generally, if C is an m-dimensional curve given by simple closed or Jordan iff it is both simple and closed. r(s), then for f : Rm → R, Z Definition 170. A point set P is connected (or more n X f (r(s))ds = lim f r(s∗i ) ∆si . (134) precisely path-connected ) iff for any two points in P there n→∞ C exists a path in P that connects them. i=1 Definition 162. The line integral of f (x, y) along a curve C (with respect to x and y) is Z n X f (x, y)dx = lim f (x∗i , yi∗ )∆xi , (135a) C n→∞ Z f (x, y)dy = lim C n→∞ Theorem 171 (Jordan curve). The complement of a simple closed curve in the plane consists of two components: the bounded one is called its interior and the unbounded its exterior ; both are open and path-connected. i=1 n X i=1 f (x∗i , yi∗ )∆yi . Definition 172. A point set P is simply-connected iff it (135b) is connected and every simple closed curve in P encloses only points that are in P. 11 Qinghai Zhang Summary of Math 1321-004 2013-MAR-18 R Theorem 173. F · dr is independent of path in D iff Definition 183. A surface is orientable if it has two H F · dr = 0 for all closed path C ⊂ D. sides. The positive side is the side to which the unit norC m m Theorem 174. Let F : R → R be a vector field that mal vector points. By convention, a simple closed surface is oriented by its unit outward normal vector. Ris continuous on an open and connected region D. If F · dr is independent of path in D, then F is conser- Definition 184. For a continuous vector field F defined C vative on D. on an oriented surface S with unit normal vector n, the 1 Theorem 175. If F = hP (x, y), Q(x, y)i is C and con- surface integral of F over S (or the flux of F across S) ∂Q is ZZ ZZ servative on D, then ∀(x, y) ∈ D, ∂P ∂y = ∂x . F · dS = F · n dS. (147) Theorem 176. If F = hP (x, y), Q(x, y)i is C 1 on an S S ∂Q open and simply-connected region D and ∂P ∂y = ∂x , then Formula 185. Consider a continuous vector field F is conservative. F = hP (x, y, z), Q(x, y, z), R(x, y, z)i and an oriented Definition 177. The surface integral of f (x, y, z) over surface S given by z = g(x, y). ZZ ZZ a surface S given by r(u, v) = hxr (u, v), yr (u, v), zr (u, v)i ∂g ∂g −Q + R dA, (148) F · dS = −P is ∂x ∂y S D ZZ m n XX f (Pij∗ )∆Sij , (142) where D is the projection of S onto the xy-plane. f (x, y, z)dS = lim m,n→∞ S where Pij∗ = i=1 j=1 Theorem 186 (Gauss’s). Let V be the interior of a simple closed surface ∂V. If F : R3 → R3 is C 1 on an open region that contains V, then ZZ ZZZ F · dS = ∇ · F dV. (149) r(u∗i , vj∗ ). Formula 178. If a smooth surface S is given by r(u, v) = hxr (u, v), yr (u, v), zr (u, v)i, then ZZ ZZ ∂r ∂V V ∂r f (x, y, z)dS = f (r(u, v)) × dA, (143) ∂u ∂v Theorem 187 (Green’s). Let D be the interior of S D a positively-oriented, piecewise-smooth, simple closed where D is the integral domain in the uv-plane. curve ∂D in the plane. If F = hP (x, y), Q(x, y)i is C 1 Formula 179. If a smooth surface S is given by on an open region that contains D, then z = g(x, y), then I ZZ ∂Q ∂P ZZ P dx + Qdy = − dA. (150) ∂x ∂y f (x, y, z)dS (144) ∂D D S s Equivalently, if G = hP (x, y), Q(x, y), 0i, then 2 ZZ 2 ∂z ∂z I ZZ = f (x, y, g(x, y)) + + 1 dA, ∂x ∂y G · dr = (∇ × G) · kdA. (151) D ∂D where D is the projection of S onto the xy-plane. Definition 180. The unit outward normal vector of a closed 2D curve r(t) = hx(t), y(t)i is n(t) = −N(t), D Equivalently, if H = hQ(x, y), −P (x, y)i, then I ZZ H · n ds = ∇ · H dA, ∂D (145) (152) D where n is defined in (145). where N(t) is the principal unit normal vector defined in Formula 188. The area of a region D bounded by a Definition 94. simple closed curve ∂D is given by Definition 181. A simple closed curve is positively oriI I I 1 ented if an observer traversing the curve always has the A(D) = xdy = − ydx = xdy − ydx. 2 ∂D interior of the curve on her left side. ∂D ∂D (153) Definition 182. The unit normal vector of a surface Theorem 189 (Stokes’). Let S be an oriented piecewiser(u, v) is ∂r ∂r smooth surface with its boundary ∂S a simple, closed, × ∂v n(u, v) = ∂u . (146) piecewise-smooth curve. If F : R3 → R3 is C 1 on an ∂r ∂r × ∂u ∂v open region that contains S, then n(u, v) is called the unit outward normal vector if the Z ZZ surface r(u, v) is simple closed and n(u, v) points from F · dr = (∇ × F) · dS. (154) its interior to its exterior. ∂S S 12