MA1521 Calculus for • Common integrals: Z • Notable formula:

Computing y −→ y dx Z
A cos θ + B sin θ
dθ
sin x −→ − cos x + C cos θ + sin θ
Z A+B A−B
Limits cos x −→ sin x + C 2 (cos θ + sin θ) + 2 (cos θ − sin θ)
= dθ
tan x −→ − ln |cos x| + C cos θ + sin θ
• Maxima: f (c) is local maximum if f (c) ≥ f (x) for x Z
A+B
Z
A − B cos θ − sin θ
near c (note that inequality is non-strict) cot x −→ ln |sin x| + C = dθ + dθ
2 2 cos θ + sin θ
sec x −→ ln |sec x + tan x| + C
• Differentiability: f is differentiable at a if f 0 (a) exists A+B A−B
csc x −→ − ln |csc x + cot x| + C = θ+ ln (cos θ + sin θ) + C
f is differentiable (in its domain) if f is differentiable at 2 2
every point in its domain
f is differentiable at point a =⇒ f is continuous at a • Trigonometric formulae: 1 + cot2 x = csc2 x
• Trigonometric formulae:
(but not the converse) d n
• Power rule: x = nxn−1
• Critical Point: c is a critical point if c is an interior dx
Pythagorean Double angle
point in the domain and either f 0 (c) = 0 or f 0 (c) does 0 0
• Product rule: (f g) (x) = f (x)g(x) + f (x)g (x) 0
not exist (A critical point may not be a local extremum) sin 2x = 2 sin x cos x
 0 sin2 x + cos2 x = 1
f f 0 (x)g(x) − f (x)g 0 (x)
• Inflection Point: c is an inflection point if concavity • Quotient rule: (x) = 2 tan2 x + 1 = sec2 x cos 2x = cos2 x − sin2 x
changes at c g [g(x)]
cot2 x + 1 = csc2 x = 2 cos2 x − 1
(c is an inflection point =⇒ f 00 (c) = 0, but the
• Chain rule: (f ◦ g)0 (x) = f 0 (g(x))g 0 (x) = 1 − 2 sin2 x
converse is not always true)
dy dy dx 2 tan x
sin (x) • Parametric differentiation: = ÷ Triple angle tan 2x =
• lim =1 dx dt dt 1 − tan2 x
x→0 x sinh 2x = 2 sinh x cosh x
• Implicit differentiation: Differentiate both sides of an sin 3x = 3 sin x − 4 sin3 x
• L’Hôpital’s Rule: cosh 2x = 2 cosh2 x − 1
f (x) f 0 (x) equation containing x and y, then solve the resulting cos 3x = 4 cos3 x − 3 cos x
f (a) = g(a) = 0 or ∞ =⇒ lim = lim 0 dy 2 tanh x
x→a g(x) x→a g (x) equation for the term. tanh 2x =
use ln (. . . ) for other forms: ∞ − ∞, 1∞ , ∞0 , 00 dx 1 + tanh2 x
Z Z
• Integration by parts: u dv = uv − v du
Differentiation & Integration Try to differentiate in this order (highest to lowest Sum of angles
1 priority):
• Common derivatives: loga x −→ x ln a sin(α ± β) = sin α cos β ± cos α sin β
y −→
dy sin−1 x −→ √ 1
ln x xn ex , e−x sin x, cos x
1−x2 cos(α ± β) = cos α cos β ∓ sin α sin β
dx
cos−1 x −→ − √1−x 1
tan α ± tan β
sin x −→ cos x 2
d
Z x
tan(α ± β) =
cos x −→ − sin x tan−1 x −→ 1 • f (t) dt = f (x) 1 ∓ tan α tan β
1+x2 dx a
−1 1 cot α cot β ∓ 1
tan x −→ sec2 x cot x −→ − 1+x 2 cot(α ± β) =
d
Z g(x) cot β ± cot α
cot x −→ − csc x2 sec −1
x −→ √1
|x| x2 −1
• f (t) dt = f (g (x)) g 0 (x)
dx a
sec x −→ sec x tan x csc−1 x −→ − |x|√1x2 −1 √ Linear combination of sine and cosine
csc x −→ − csc x cot x • Integral with R2 − x2 : Sub x = R sin θ or x = R cos θ
x x
x −→ x (ln x + 1) √
ax −→ ax ln a
p
• Integral with R2 + x2 : Sub x = R tan θ a sin θ + b cos θ = a2 + b2 sin (θ + arctan2 (b, a))
1
 ∞
Series 1
=
X
(−1)n x2n = 1 − x2 + x4 − x6 + · · ·
Partial Differentiation
1 + x2 
• Convergence: A sequence {an } is convergent if for n=0 ∂f 
∞
(−1)n−1 xn x2 x3 x4 • Partial derivative wrt x – denoted by fx (a, b) or
some fixed L ∈ R, we have lim an = L ∂x (a,b)
X
n→∞ ln (1 + x) = =x− + − + ···
n 2 3 4
Otherwise, the sequence is divergent n=1
∞ ∂2f
X (−1)n 2n+1 x3 x5 x7 • fxy = (fx )y = ∂y∂x
1−rn tan−1 x = x =x− + − + ···
• Partial sum of geometric series: sn = a , 2n + 1 3 5 7
1−r n=0
∞ • fxy = fyx (when f is continuous in the neighbourhood)
r 6= 1 1 X
sn converges ⇐⇒ |r| < 1 = nxn−1 = 1 + 2x + 3x2 + 4x3 + · · ·
(1 − x)2 • To check if f (x, y) has partial derivatives of all orders,
n=1
∞ check if fxy = fyx
• Ratio test for convergence of series: 1 1X 1
) converges = n(n − 1)xn−2 = (2 + 6x + 12x2 + · · · )
if ρ < 1 (1 − x)3 2 n=2 2
( n
• Chain rule: For z = f (x, y) and x = x(t), y = y(t):
 
 an+1  X 
lim   = ρ, {sn } = ai diverges if ρ > 1 ∞
n→∞ an  1 X
unknown if ρ = 1 = (−1)n (n + 1)xn = 1 − 2x + 3x2 − 4x3 + · · · dz ∂f dx ∂f dy

i=1
(1 + x)2 = +
n=0 dt ∂x dt ∂y dt
• Radius of convergence: for −∞ < x < +∞
The value of r, where ρ < 1 =⇒ |x − a| < r ∞ For z = f (x, y) and x = x(s, t), y = y(s, t):
(for power series — may be 0, some positive real, or ∞)
X (−1)n x2n+1 x3 x5 x7
sin x = =x− + − + ···
n=0
(2n + 1)! 3! 5! 7! ∂z ∂f ∂x ∂f ∂y
∞  = +
X 1 converges if p > 1 ∞
(−1)n x2n x2 x4 x6 ∂s ∂x ∂s ∂y ∂s
• p-series: cos x =
X
=1− + − + ···
n=1
np diverges if 0 ≤ p ≤ 1 (2n)! 2! 4! 6!
n=0
∞ • Directional derivative: For unit vector u = u1 i + u2 j:
∞ X xn x2 x3
ex = =1+x+ + + ···
X n
• Power series about x = a: f (x) = cn (x − a) n! 2! 3!
n=0 Du f (a, b) = fx (a, b)u1 + fy (a, b)u2
n=0

• Differentiation & integration of power series: • Gradient vector: ∇f (a, b) = fx (a, b)i + fy (a, b)j
∞ • Deriving more Maclaurin series:
∇f (a, b) · u = Du f (a, b)
X n−1
f 0 (x) = ncn (x − a) Substitution: f (x) = g(x) =⇒ f (h(x)) = g(h(x))
n=1 Multiplication: f (x) = g(x) =⇒ h(x)f (x) = h(x)g(x) Direction of ∇f (a, b) is the steepestpdirection, maximum
radius of convergence is unchanged by differentiation Differentiation: f (x) = g(x) =⇒ 0 0 value of Du f (x, y) is k∇f (x, y)k = fx (a, b)2 + fy (a, b)2
Z X∞
(x − a)
n+1 Z f (x) = g (x)
x Z x
f (x)dx = cn +C Integration: f (x) = g(x) =⇒ f (x)dx = g(x)dx • Maxima: f (a, b) is local maximum if f (a, b) ≥ f (x1 , y1 )
n+1 0 0
n=0 for all points (x1 , y2 ) near (a, b) (note that inequality is
radius of convergence is unchanged by integration ... and other usual operations on functions
non-strict)
∞
X f (k) (a) k
• Taylor series: f (x) = (x − a) • Taylor polynomial: • Critical point: (a, b) is a critical point if fx (a, b) = 0
k! and fy (a, b) = 0 (not all critical pts are min/max points)
k=0 nth order Taylor polynomial: Pn (x) := terms until (and
• Maclaurin series: Taylor series with a = 0 including) xn , use for approximation A maximum point might not be a critical point if f is
remainder of order n: Rn (x) := remaining terms not smooth
• Common Maclaurin series: error := absolute value of remaining terms = |Rn (x)|
for −1 < x < 1 • Discriminant: D = fxx (a, b)fyy (a, b) − fxy (a, b)2
∞ If D > 0 and fxx (a, b) > 0 then (a, b) is a local minimum
1 X
f (n+1) (c) If D > 0 and fxx (a, b) < 0 then (a, b) is a local maximum
= xn = 1 + x + x2 + x3 + · · · • Taylor’s Theorem: Rn (x) = (x − a)
n+1
1 − x n=0 (n + 1)! If D < 0 then (a, b) is a saddle point
∞
1 X where x ≤ c ≤ a (when x ≤ a) or a ≤ c ≤ x (when a ≤ x) If D = 0 then no conclusion can be drawn
= (−1)n xn = 1 − x + x2 − x3 + · · ·
1 + x n=0 This provides an upper bound for the error term For finding max/min, fyy may be used in place of fxx
2
Ordinary Differential Equations • Superposition principle: If y1 and y2 are solutions • Logistic growth model:
then c1 y1 + c2 y2 is also a solution (i.e. the solution set is B := per capita birth rate (constant)
n
X a subspace) sN := per capita death rate (linear to population)
• Definition: ai (x)y (i) (x) = F (x) dN
i=0 = (B − sN )N =⇒
where ai (x) and F (x) are functions of x dt
• Dimension: The solution space has dimension 2, so B B
and y (i) (x) is the i th derivative of y w.r.t. x finding 2 linearly independent solutions is sufficient to N (t) = = s
s + ( NB(0) − s)e−Bt
 
B 1
obtain the general solution 1+ s· − 1 e−Bt
N (0)
• Separable equations: Separate and integrate both
sides B − sN (t) > 0 ∀t =⇒ Bs > N (t)
=⇒ smaller than sustainable population
dx • Guessing: Obtain 2 lin. indep. solutions by guessing B − sN (t) = 0 ∀t =⇒ Bs = N (t)
• Exponential decay: = kx =⇒ x(t) = x(0)ekt
dt =⇒ sustainable population (equilibrium)
ln 2
x(0) is the initial val., k = − where τ is the half-life B − sN (t) < 0 ∀t =⇒ Bs < N (t)
τ • Constant p(x) and q(x):
=⇒ larger than sustainable population
d2 y dy
• Exponential cooling/heating: Let +A + By = 0, then solutions have form B
dx dx2 dx Population always tends to — “carrying capacity”
= k(x − x0 ) =⇒ x(t) − x0 = (x(0) − x0 ) ekt λx
y = e for some value λ s
dt
x(0) is the initial value, x0 is the target value By substitution, λ2 + Aλ + B = 0, solve for λ = λ1 , λ2
• Harvesting growth model:
Two distinct real roots: y = c1 eλ1 x + c2 eλ2 x
e −e x −x B := per capita birth rate (constant)
• Hyperbolic functions: sinh(x) = Two repeated (real) roots: y = c1 eλx + c2 xeλx √
sN := per capita death rate (linear to population)
2 Two distinct complex roots: If λ1 or λ2 = a + b −1
ex + e−x ex − e−x E := harvest rate
cosh(x) = tanh(x) = x then y = eax (c1 cos bx + c2 sin bx) dN
2 e + e−x = (B − sN )N − E = BN − sN 2 − E
Given λ1 , λ2 we can recover A = −(λ1 + λ2 ), B = λ1 λ2 dt
• Reduction to separable form: If ODE contains Solve −sN 2 + BN − E = 0 for equilibrium solutions
2
fraction where the degree of terms are equal, then we B 2 − 4sE > 0 =⇒ B4s > E =⇒ two equilibriums
can substitute: =⇒ β1 < β2 < Bs and β1 + β2 = Bs
dy
=
y 2 − x2 y
= −
x
=⇒ x
dv
+v =v−
1 Non-homogeneous linear 2nd order ODEs β2 is stable but not β1
2
dx xy x y dx v B − 4sE = 0 =⇒ B4s = E =⇒ one equilibrium
2
y d2 y dy B
where v = + p(x) + q(x)y = f (x) =⇒ β = 2s
x dx2 dx 2
B 2 − 4sE < 0 =⇒ B4s < E =⇒ no equilibrium
dy Find a particular solution yp for this equation, and find the
• Substitution: Substitute u = f (x, y) to expr. and general solution yh for the homogeneous version (f (x) = 0),
dx
then the general solution for this equation is y = yh + yp Useful Sequences
dy
• Linear 1st order ODEs: + p(x)y = q(x)
dx
Prime Numbers
Z R
=⇒ yf (x) = q(x)f (x)dx where f (x) := e p(x)dx
2 3 5 7 11 13 17 19 23 29
Mathematical Modelling 31 37 41 43 47 53 59 61 67 71
dy
• Bernoulli equation: + p(x)y = q(x)y n 73 79 83 89 97 101 103 107 109 113
dx • Malthusian growth model:
dz 127 131 137 139 149 151 157 163 167 173
=⇒ + (1 − n)p(x)z = (1 − n)q(x) where z = y 1−n B := per capita birth rate (constant) 179 181 191 193 197 199 211 223 227 229
dx
... which is a linear 1st order ODE D := per capita death rate (constant) 233 239 241 251 257 263 269 271 277 ···
dN
= (B − D)N =⇒ N (t) = N (0)e(B−D)t
Homogeneous linear 2nd order ODEs dt 1521 = 32 × 132
B > D =⇒ population explosion 2017 is prime
d2 y dy B = D =⇒ stable
+ p(x) + q(x)y = 0 B < D =⇒ extinction
dx2 dx
3